WO2009074089A1

WO2009074089A1 - Image or image mode coding or decoding method, apparatus and image coding and decoding ststem

Info

Publication number: WO2009074089A1
Application number: PCT/CN2008/073277
Authority: WO
Inventors: Jianhua Zheng
Original assignee: Huawei Technologies Co., Ltd.
Priority date: 2007-11-30
Filing date: 2008-12-01
Publication date: 2009-06-18
Also published as: CN101453642B; CN101453642A

Abstract

An image and image mode coding and decoding method, apparatus and an image coding and decoding system. The image coding method comprises making sure the image mode value corresponding to the image information according to the image information of image to be encoded; coding the image information to get the coded code stream with the image mode value; coding the image mode value and logging the coded image mode value in the coded code stream. The image decoding method comprises parsing the code stream of image to be decoded to get the image mode value of image to be decoded; decoding the code stream of image to be decoded with the image mode value; outputting the decoded image.

Description

BACKGROUND OF THE INVENTION 1. Field of the Invention Priority of the application, the entire contents of which are incorporated herein by reference.

Technical field

The present invention relates to the field of image coding and decoding technologies, and in particular, to an image and image mode coding and decoding method and apparatus, and an image coding and decoding system. Background technique

In the video image coding and decoding technology, the input signal is usually an image sequence, and the image sequence is usually divided into one image and one image, which is also referred to as an image frame input to the encoder. Therefore, the image is a processed signal of the codec. Basic unit.

In practical applications, the characteristics of the captured images are significantly different because of the different shooting conditions of the images. Different shooting conditions will directly affect the quality of the captured image; while the shooting scene is different, the characteristics of the captured image will change greatly. In order to obtain better image quality, the photographer and camera equipment are higher. Claim. The photographer is required to adjust the aperture, shutter, focus, metering method, and flash of the camera according to the subject and the shooting scene. In order to solve the impact on the quality of the captured image due to different shooting conditions and shooting conditions, digital cameras and other manufacturers have added several scene modes to the digital camera, and pre-adjust the aperture, shutter, focal length, metering method and flash. The value of the parameters makes it easier to take high-quality photos, even for those who are not experienced enough to take digital photos with certain quality assurance. At present, there are four or five scene modes in digital cameras, and there are twenty or thirty kinds of scenes. For example: When using landscape mode, the digital camera will adjust the aperture to the minimum to increase the depth of field, so that the photos get the clearest effect. ; When using portrait mode, the digital camera can adjust the hue, contrast or softening effect to make the portrait of the person perform a stronger skin tone effect; when using the night scene mode, the digital camera extends the shutter exposure time to ensure that the photo is fully exposed, and the photo will also be Brighter; When using dynamic mode (sport mode), the digital camera adjusts the shutter speed to a faster or higher ISO sensitivity for shooting high-speed moving objects. Scene mode in a digital camera The format corresponds to the optical characteristics of the digital camera, such as aperture, shutter, focus, metering and flash. These parameters can be selected by the photographer to get better results when shooting. However, for the video sequence, the sequence of video images obtained by the photographer after the shooting is completed will be fixed, and these video image sequences are used as input signals of the video image encoder. For video image encoders, it is generally impossible to determine the shooting conditions or shooting conditions, and it is not possible to control the optical characteristics of the shooting device.

In addition, the video image encoder encodes the pixel value of the image, even if a digital camera-like optical characteristic scene mode is used in shooting, but the sequence characteristics of shooting in different scenes still vary greatly, which encodes the video image. The encoding performance of the device has a great influence, such as: movie clips in different scenes. And the optical characteristic parameters of the shooting device are not directly related to the video image encoder. These optical characteristic parameters cannot be used in the video image encoder.

Currently, in the joint photographic experts group (Jpeg), moving picture experts grou (MPEG) specification 1 (Mpegl), Mpeg2, Mpeg4, H.264, AVS and other image coding standards or moving images In the coding standard, the image or video image sequence is not distinguished according to the image scene, that is, the entire video image sequence is encoded according to a unified coding method, and the coding parameters are set before encoding. However, the input video image sequence differs greatly in the characteristics of the image due to the difference in the shooting scene, and the encoding method without distinguishing the image scene cannot be adapted to the change of the video image.

In film and television applications, a 90-minute movie may include dozens of clips and scenes. Different scenes, lighting, background, movement of people or objects, etc., will result in different video sequences. The content characteristics of scenes or clips are different. It is obviously not suitable for the change of video content by using the same encoding method for the sequences of these different characteristics.

In video surveillance applications, there are many application scenarios for surveillance cameras, such as: indoor monitoring, outdoor monitoring, infrared monitoring, etc. Even in the same monitoring scene, the characteristics of the monitoring sequence are changed significantly due to changes in ambient lighting, for example. : Day, night, rainy days, rain and snow, etc. The application change of these monitors or the change of ambient illumination has a great influence on the coding performance of the video encoder. Using the same coding method to encode the sequences of these different scenes will result in low coding performance of the video encoder, reducing coding efficiency or The severe imbalance of coded video quality in different scenarios leads to a decrease in the subjective quality of image coding. Summary of the invention

Provided are an image coding method, apparatus and system, an image decoding method, apparatus and system, an image mode coding method and apparatus, an image mode decoding method and apparatus, and an image coding and decoding system, which may be In the video image encoding and decoding process of scene change, the video image encoding and decoding compression performance of different scenes is improved.

An image encoding method, comprising:

Determining an image mode value corresponding to the image information according to image information of the image to be encoded; encoding the image information by using the image mode value to obtain an encoded code stream; encoding the image mode value, and encoding The latter image mode value is written into the encoded code stream.

An image decoding method includes:

Parsing the code stream of the decoded image to obtain an image mode value of the image to be decoded; decoding the code stream of the image to be decoded by using the image mode value, and outputting the decoded image.

An image mode encoding method includes:

Determining an image mode value of the image to be encoded;

The image mode value is encoded and the encoded image mode value is written into the encoded code stream.

An image mode decoding method includes:

Parsing and determining an image mode value of the currently decoded image from the image code stream;

An image mode is determined based on the image mode value.

An image encoding device comprising:

a first module, configured to determine an image mode value corresponding to the image information according to image information of the image to be encoded;

a second module, configured to encode the image information by using the image mode value to obtain a coded code stream;

And a third module, configured to encode the image mode value, and write the encoded image mode value into the encoded code stream. An image decoding device includes:

a fourth module, configured to parse a code stream of the image to be decoded, to obtain an image mode value of the image to be decoded;

And a fifth module, configured to decode the code stream of the image to be decoded by using the image mode value, and output the decoded image.

An image mode encoding device includes:

a sixth module, configured to determine a scene mode value of the image to be encoded;

And a seventh module, configured to encode the image mode value, and write the encoded image mode value into the encoded code stream.

An image mode decoding device includes:

And an eighth module, configured to parse and determine an image mode value of the currently decoded image from the image code stream; and a ninth module, configured to determine an image mode according to the image mode value.

An image coding system comprising:

And an obtaining module, configured to determine an image mode value corresponding to the image information according to image information of the image to be encoded;

a first encoding module, configured to encode the image information by using an image mode value determined by the acquiring module to form an encoded code stream;

a second encoding module, configured to encode an image mode value determined by the acquiring module, and store the image mode value in the encoded code stream;

And a transmission module, configured to output the encoded code stream carrying the image mode value.

An image decoding system, comprising:

a parsing module, configured to parse a code stream of the image to be decoded, to obtain an image mode value of the image to be decoded;

a decoding module, configured to decode the code stream according to an image mode value obtained by the parsing module;

And an output module, configured to output the decoded image of the decoding module.

An image coding and decoding system, comprising:

An acquiring module, configured to determine an image mode value corresponding to the image information according to image information of an image to be encoded; a first encoding module, configured to encode the image information by using an image mode value determined by the acquiring module to form an encoded code stream;

a transmission module, configured to output the coded code stream carrying the image mode value;

It can be seen from the above technical solution that the embodiment of the present invention incorporates an encoding method of image mode characteristics in video coding, and provides a certain encoding and decoding of different scene images by encoding video images according to image mode values of different video image sequences. The ability to adapt the scene to improve the codec compression performance in different scenarios. DRAWINGS

Figure 1-a is a schematic diagram of a sequence of video images;

Figure 1-b is a schematic diagram of a video image sequence divided into different scene image sequences;

2 is a schematic diagram of a frame-type flow reference of an embodiment of an image encoding method according to the present invention;

3 is a schematic diagram of a frame-type flow reference of an embodiment of an image decoding method according to the present invention;

4 is a schematic flow chart of an embodiment of an image encoding method according to the present invention;

FIG. 5 is a schematic flowchart diagram of an embodiment of an image decoding method according to the present invention; FIG.

6 is a schematic flow chart of an embodiment of an image mode encoding method according to the present invention;

Figure Ί is a schematic flowchart of an embodiment of an image mode decoding method according to the present invention;

8 is a schematic diagram of a frame-type flow reference using a typical video image encoder corresponding to transform, scaling/quantization, and entropy coding in an embodiment of an image encoding method according to the present invention;

FIG. 9 is a schematic diagram of a frame-type flow reference of a typical video image decoder corresponding to an inverse transform, a scaling/inverse quantization, and an entropy decoding according to an embodiment of an image decoding method according to the present invention; FIG.

Figure 10-a is a schematic diagram of a scene image band distribution template in a default scene mode; Figure 10-b is a schematic diagram of a scene image band distribution template in an infrared scene mode; Figure 10-C is a schematic diagram of a scene image band distribution template in a night scene mode;

Figure 10-d is a schematic diagram of a scene image band distribution template in a moving scene mode;

11 is a schematic structural diagram of Embodiment 1 of an image coding system according to the present invention;

12 is a schematic structural diagram of Embodiment 2 of an image coding system according to the present invention;

13 is a schematic structural diagram of Embodiment 1 of an image decoding system according to the present invention;

FIG. 14 is a schematic structural diagram of Embodiment 2 of an image decoding system according to the present invention. detailed description

In order to facilitate the understanding of the specific embodiments of the present invention, the specific implementation process of the method of the embodiments of the present invention will be described in detail below with reference to specific examples.

In order to better describe the method of the present invention, the sequence of video images taken under different conditions is defined as follows.

Definition 1: [Scenario of video sequence image] For consecutively captured image sequences of the same camera or the same object, define several contacts of the same camera from different angles or different cameras to the same location or the same object. The sequence of images is a scene. There are large similarities between image features of the same scene, and image sequences of different scenes are quite different in image features.

As shown in Figure 1-a, it is a schematic diagram of a sequence of video images, containing a series of images, and different images are taken in different scenes. There is a great similarity between the features of the individual images labeled A; the features between the individual images labeled B also have a large similarity; but between the images labeled A and B. Feature gap ^ 艮艮 big. As shown in FIG. 1-b, the video image sequence is divided into different scene image sequences, that is, the image with similar image feature similarity, and the image sequence belonging to the mark A is divided into the scene A; similarly, the scene B can be obtained. Sequence, scene C sequence, such that the input image sequence of the image encoder consists of several different sequence of scenes.

For a sequence of video images, the scene also appears as a set of images before or after the switching of the image content of the same video image sequence, or a video footage segment. Therefore, a video sequence can also be divided into several scene segments or video footage segments, each of which is composed of a sequence of images containing the scene or lens. For convenience, the following description of the present invention will A sequence of video scenes or video footage is not distinguished, and is collectively referred to as a sequence of scenes.

Obviously, there is a great similarity between the features of each image in the same scene sequence, and the image features contained between different scene sequences are quite different.

Definition 2: [Scene mode of the image] That is, the scene mode of the scene sequence image, the image features of the same scene have greater similarity, and the image sequences of different scenes have different characteristics on the image features, so , video images in the same scene or shot can be represented by a cluster. The scene mode is a set of parameter values or sets of parameters corresponding to different scene image features, which represent common image features of the sequence of images contained in the scene or shot.

The image involved in the embodiment of the present invention includes a still image, a moving image, a residual image of two adjacent images of the moving image, a target image obtained by computing an arbitrary image in the moving image, and the like.

An image scene mode in the embodiment of the present invention may include two types of values:

The first type, scene mode identification (scene mode identification codeword or scene mode index).

An image scene used to identify different features in an image or video stream. Different image scenes are identified by unique symbols or by an index. The current scene mode can be identified by the identification value or the index value, that is, in the code stream. When the image scene is switched, only update or switch the scene mode identification value.

In the second category, the scene mode parameter (the form of the scene mode parameter or the scene mode parameter set exists). A set of parameters or sets of parameters representing images of different scenes, the set of parameters or sets of parameters corresponding to the scene mode. In the video image encoding process, these parameters or parameter sets corresponding parameter values are used in the video encoding and decoding process as the encoding parameters of the video encoding process or the decoding parameters of the decoding process, thereby achieving different encoding and decoding performance and different image quality. Video stream.

The identifier value of the scene mode is a codeword, and the codeword uniquely corresponds to a scene mode. When there are many scene modes, an index may be used to distinguish different scene modes. A scene mode identification value corresponding to a set of scene mode parameters or parameter sets.

In addition, the above scene mode parameters can include three categories:

The first type, the feature parameter value representing the feature of the scene image obtained from the image information.

Therefore, different scene images have different feature parameter values because of different image characteristics; similar scene images, because the image features are similar, the corresponding feature parameter values are similar. The values of the feature value parameters obtained from the image information include, but are not limited to, the illumination brightness of the image scene, the illumination chromaticity of the image scene, the coefficient band division manner of the transform domain in the image coding, the coefficient band parameter value of the transform domain, and the parameter values thereof. Distribution, the way the target or background moves in the image scene.

The second type, the current image coding information obtained in image encoding or decoding.

These information are image coding information calculated by other modules in the image coding or image decoding process, such as motion vectors, or statistically obtained image coding information, such as the type and number of coding blocks.

The third category, image coding or decoding, exists and is used to control the control parameters of the encoder encoding process or the decoder decoding process.

The parameters include, but are not limited to, the size of the picture group (GOP), the coded picture quantization parameter QP, the number of reference frames, the motion vector search range, the loop filter control, the entropy coding control, and the like.

The above-mentioned scene mode parameter set is a form of one or more types of parameters or a combination thereof among the above three types of scene mode parameters. Image coding method embodiment

FIG. 4 is a schematic flowchart diagram of an embodiment of an image encoding method according to the present invention. The encoding method of this embodiment may include the following steps:

Step 401: Acquire image information of an image to be encoded, where the image information includes a pixel value of the image, a residual value, a pixel value or a residual value of the image in the transform domain, and corresponding coding information in the image coding.

Step 402: Determine a scene mode value of an image corresponding to the image according to image information of the image to be encoded. The scene mode value includes a scene mode identifier value; or includes a scene mode identifier value and a corresponding scene mode parameter value.

Step 403: Determine, according to the scene mode of the image to be encoded, whether the current scene mode being used in the encoding has changed. When the scene mode identification value changes in the scene mode, or the corresponding scene mode parameter value changes, the current scene mode may be changed.

Step 403 includes:

Step 4031: Determine whether the current scene mode is changed. If the process is changed to step 4032, otherwise go to step 4033.

Step 4032: If the current scene mode is changed, use the scene mode determined in step 402. As the current scene mode, and update the corresponding scene mode identification value, and the corresponding scene mode parameters.

Step 4033: If the current scene mode has not changed, the scene mode used in the current encoding is suitable for the image to be encoded, and does not need to be changed.

Step 404: Perform image coding on the image to be encoded by using the current scene mode. The current scene mode can participate in the image encoding process as a control parameter in image encoding. When encoding the current input image, the current scene mode may participate in or determine the coding scheme and process including, but not limited to, the following, such as image block quantization method, image block band division method, image intra prediction method, image inter prediction method, and the like. . Thereby, an image coded code stream of different image coding methods, image coding quality, image coding rate, and the like is obtained.

Step 405: Encode the scene mode value, and write the encoded scene mode value into the encoded code stream. The coding can use a fixed length code, a variable length code, an entropy coding, and the like. Encoding the scene mode values includes encoding the scene mode identification values, or encoding the scene mode parameters. The encoded result will be written to the video image encoded bitstream.

Step 406: Output the encoded result including the scene mode coded code stream and the image coded code stream.

The steps 404 and 405 have no order and can be interchanged.

In addition, the understanding of this embodiment can also be referred to FIG. 2, which is a schematic diagram of a frame-type flow reference of an embodiment of the image coding method of the present invention.

The image coding method in this embodiment integrates an encoding method according to a scene mode characteristic in video coding, and controls a coding characteristic parameter of a video coding, such as a quantization parameter QP, a frequency band division manner, a frequency band parameter size, by a scene mode in video coding, Intra prediction mode, inter prediction compensation, motion estimation, and the like. When the scene switching occurs in the coding sequence, the image coding method of the embodiment can be adapted to the image features of the transformed scene to improve the compression ratio of the image coding. Image decoding method embodiment

FIG. 5 is a schematic flowchart diagram of an embodiment of an image decoding method according to the present invention. The decoding method of this embodiment may include the following steps:

Step 501: Parsing an input image code stream to be decoded, and obtaining the decoded image. Scene mode value. The scene mode identifier value is included; or the scene mode identifier value and the corresponding scene mode parameter value are included.

Step 502: Determine, according to the parsed scene mode value of the image to be encoded, whether the current scene mode being used in the current decoding process has changed. When the scene mode identification value changes in the scene mode, or the corresponding scene mode parameter value changes, the current scene mode may be changed.

Step 502 includes:

Step 5021: Determine whether the current scene mode is changed. If the process proceeds to step 5022, otherwise go to step 5023.

Step 5022: If the current scene mode is changed, use the scene mode parsed in step 502 as the current scene mode, and update the corresponding scene mode identifier value, and the corresponding scene mode parameter value.

Step 5023: If the current scene mode has not changed, the current scene mode will continue to be used in image decoding.

Step 503: Perform image decoding on the image to be decoded by using the current scene mode value.

The current scene mode value may participate in or determine coding schemes and processes including, but not limited to, image block inverse quantization mode, image block band division method, image intraframe decoding, image interframe prediction compensation, and the like. Thereby a reconstructed image sequence of different image decoding qualities is obtained.

Step 504: Output the decoded reconstructed image.

In addition, the understanding of this embodiment can also be referred to FIG. 3, which is a schematic diagram of a frame-type flow reference of an embodiment of an image decoding method according to the present invention.

In the image decoding method of the embodiment, the scene mode information existing in the encoded code stream is parsed, and whether the scene mode in the decoding process is switched according to the current scene mode change or not, so that the decoding process switches the image sequence to the scene. Have a certain ability to adapt. And through the scene mode to participate in the process of image decoding, thereby improving the quality of the decoded reconstructed image. Image mode encoding method embodiment

FIG. 6 is a schematic flowchart diagram of an embodiment of an image mode encoding method according to the present invention. In this embodiment, the scene mode is taken as an example, and the corresponding scene coding method may include the following steps: Step 601: Enter a current scene mode value to be encoded, where the above scene mode value is used to encode the current image.

Step 602: Determine a scene mode identifier value according to the scene mode value. Each scene mode contains a unique scene mode identification value (identification codeword or index value) that is used to distinguish between different scene modes.

Step 603: Encode the scene mode identifier value, and write the encoded scene mode value into the encoded code stream.

Step 604. As mentioned above, the scene mode has two types of values. Each scene mode may include a second type of value, that is, a scene mode parameter value, in addition to the first type of value, that is, the scene mode identification value. Determine whether the scene mode parameter value needs to be encoded according to whether the current scene mode contains the scene mode parameter value. If the scene mode parameter is included, go to step 605, otherwise go to step 607.

Step 605: Determine, by the scene mode identifier value, a scene mode parameter value.

Determining the scene mode parameter values includes:

Determine the type of the scene mode parameter, that is, the scene mode parameter contains one or several of the three types of mode parameters. Such as band distribution parameters, coded image quantization parameter QP, illumination brightness of image scene, illumination chromaticity of image scene, band division method, noise intensity, noise type, etc.

Determine the value of the scene mode parameter, that is, determine the value of the different types of scene mode parameters. For example, the band distribution parameter value, the coded image quantization parameter QP value, the scene illumination brightness level, the scene illumination chromaticity size, the band division type, the noise intensity level, the noise type value, and the like.

Step 606: Encode the scene mode parameter value, and write the encoded scene mode parameter value into the encoded code stream.

Step 607: Output a code stream that includes the result of the scene mode value encoding.

In this embodiment, the encoding of the image scene mode is implemented by writing the encoded scene mode identification value and the scene mode parameter value into the encoded code stream.

Image mode decoding method embodiment

FIG. 7 is a schematic flowchart diagram of an embodiment of an image mode decoding method according to the present invention. The image mode decoding of this embodiment may include the following steps:

Step 701: Determine a code stream header structure from the coded code stream. Step 702: Parse and obtain a current scene mode identifier value from the code stream header structure. The scene mode value is stored in the stream header structure, including the scene mode identifier value, or includes the scene mode identifier value and the scene mode parameter value. The parsed stream header structure may include one or more of the following hierarchical stream header structures, namely

Sequence header or sequence extension header;

a group header or a group extension header;

Image header or image extension header;

Strip group (set) head or strip group (set) extension header;

Strip head or strip extension head.

Wherein, when the image scene mode (the current scene mode identifier value) exists in the sequence header or the sequence extension header, the scene mode of the current decoding sequence is obtained by parsing; when the image scene mode exists in the group header or the group extension header, Parsing the scene mode of the current decoded picture group; when the image scene mode exists in the image header or the image expansion header, parsing the scene mode of the current decoded image; when the image scene mode exists in the strip group (set) header or strip When the group (set) is extended in the header, the scene mode of the current decoded stripe group (set) is parsed; when the image scene mode exists in the strip header or the stripe extension header, the scene mode of the current decoded stripe is parsed;

In a code stream header structure of a certain level, the default mode of the scene mode when the scene mode does not exist is; if the image scene mode does not exist in the code stream header structure of the same level, the current scene mode defaults to the same level. The scene mode in the previous stream header structure is the same. For example, if there is no image scene mode in the current decoded image header, the decoded image scene mode defaults to the same scene mode as in the previous decoded image header.

If the image scene mode does not exist in the same stream header structure, the current scene mode defaults to the same scene mode as the previous level stream header structure. For example, if there is no scene mode in all currently decoded image headers, the current scene mode defaults to the image level upper layer and the scene mode in the image group header structure; if there is no scene mode in the image group hierarchy, the current scene The mode defaults to the scene mode in the upper layer of the image group, the sequence header or the sequence extension header structure.

Step 703: Determine, by using the scene mode identifier value obtained in step 702, a current scene mode to be used in the image decoding process.

Step 704: Determine, according to the scene mode identifier value parsed in the code stream, whether the current scene mode is A change has occurred, and if no, go to step 708 to directly output the current scene mode; if yes, go to step 705.

If the scene mode does not exist in the encoded stream, or the parsed scene mode identifier does not change, indicating that the scene mode has not changed, the scene mode parameter value may not be analyzed. The default method of the current scene mode parameter is the same as the default mode of the scene mode.

If the scene mode identifier is changed in the code stream, it indicates that the scene mode is switched. Then, you need to continue to parse the scene mode parameter value in the stream, and update the current default value with the parsed value as the current scene mode parameter value.

Step 705: Determine, according to the code stream header structure, whether the scene mode includes a scene mode parameter value, and determine whether to continue to parse the scene mode parameter value. If not, go to step 708 to output the current scene mode. If there is a scene mode parameter, continue to parse the scene mode parameter value, and go to step 706.

Step 706: Determine, by the scene mode identifier value, a scene mode parameter value.

Wherein determining the scene mode parameter value comprises determining:

The type of the scene mode parameter corresponding to the encoding end is determined, that is, the scene mode parameter includes one or more of the three types of mode parameters. Such as band distribution parameters, coded image quantization parameter QP, illumination brightness of image scene, illumination chromaticity of image scene, band division mode, noise intensity, noise type, etc.

And determining the values of these scene mode parameters, ie determining the values of different types of scene mode parameters. For example, the band distribution parameter value, the coded image quantization parameter QP value, the scene illumination brightness level, the scene illumination chromaticity size, the band division type, the noise intensity level, the noise type, and the like.

Step 707: Decode a set of scene mode parameter values (or parameter values) in the code stream. Step 708: Output the determined scene mode, and the output result is used to participate in an image decoding process. In this embodiment, the scene mode identification value and the scene mode parameter value are decoded to determine a corresponding scene mode, thereby realizing decoding of the image scene mode.

The specific codec method in the embodiment of the image mode encoding method and the decoding method of the present invention described above will be described below with reference to specific examples.

1. Encoding method and decoding method for scene mode identification value and scene mode parameter value

Encoding end: Encoding mode can include but not limited to three methods, namely direct coding, indirect coding and entropy coding. among them, Direct encoding, that is, writing image scene mode identification values and/or scene mode parameter values directly into a video or image encoded code stream;

Indirect encoding, which converts the transformed values of image scene mode identification values and/or scene mode parameter values into a video or image encoded code stream. The above transformation methods include, but are not limited to, addition, subtraction, multiplication, division, AND, OR, and other mathematical or logical operations;

The entropy coding, that is, the image scene mode identification value and/or the scene mode parameter value, or the conversion value of the two, is entropy encoded and written into the video or image coded code stream. Entropy coding methods include, but are not limited to, fixed length code, variable length code, binarized adaptive coding, Huffman coding, run length coding, context-based variable length coding CAVLC, context-based arithmetic coding CABAC, and the like.

Decoding end: Decoding method uses the corresponding decoding method with the encoding end

If the encoding end uses direct encoding, the code stream value directly read from the code stream is the resolution result value of the scene mode identifier and its scene mode parameter.

If the encoding end uses indirect encoding, the code stream value directly read from the code stream is obtained by the inverse transform method corresponding to the encoding end to obtain an image scene mode identifier and an analysis result value of the scene mode parameter. The inverse transform method described above includes, but is not limited to, addition, subtraction, multiplication, division, AND, OR, and other mathematical or recurring operations.

If the encoding end uses entropy coding, the code stream value directly read from the code stream is obtained according to the entropy decoding method corresponding to the encoding end, and the image scene mode identifier and the resolution result value of the scene mode parameter are obtained. The above entropy decoding methods include, but are not limited to, fixed length code decoding, variable length code decoding, decoding of binary adaptive decoding methods, Huffman coding, run length coding, context-based variable length coding CAVLC, context-based arithmetic coding. CAB AC, etc.

Specifically, it is assumed here that the scene pattern that has been determined in advance is 2 ^N , N is a positive integer, assuming N = 4, and the scene mode is represented in the form of an index value.

The scene mode index value may be binary code. In this embodiment, the scene mode index value is represented by a 4-bit binary code. A comparison table of generated scene patterns and index values, as shown in Table 1.

Comparison table between scene mode and index value

0010 scene mode 2

1111 Scene Mode 15

At this time, the scene mode index is represented by U(N), which is a binary integer of N bits.

In actual applications, specific scene modes, for example, can be:

Indoor scene mode, outdoor scene mode, infrared mode, day mode, night mode, mode, rain and snow mode, image gradient mode, noise mode, distant mode, and the like.

The above index value may also represent a scene mode index value using a 1-Mbit binary variable length code. A comparison table of generated scene mode binary codes and index values can be shown in Table 2.

Scene mode binary code and index value comparison table 2

2. The method of writing the scene mode into the code stream and parsing from the code stream

Encoding end: Writing the scene mode to the code stream includes writing the scene mode identification value to the code stream, and when the scene mode parameter is present, further including writing the scene mode parameter value to the code stream.

Decoding end: Parsing the scene mode from the code stream includes parsing the scene mode identifier value from the code stream. When the scene mode parameter value is present, the method further includes parsing the scene mode parameter value from the code stream.

According to different levels of the code stream, the scene mode can exist in different code stream header structures.

For example, a sequence header or a sequence extension header containing a scene mode is as follows, and the code stream syntax of the sequence header is shown in Table 3.

Code stream syntax for sequence headers containing scene patterns

Sequence— headerO { descriptor sequence- start— code f(32)

If ( profile- id)

Scene— mode u(5) If(scene_mode==night_mode) {

Nosie—intensity se(v)

}

Scene—mode is N bits, assuming N = 5, allowing 32 common scene modes to be defined. Here sequence- header() can also be replaced by the sequence 歹 'J extension header sequence- extention-header().

In the above example, the current scene mode index scene_mode is represented by a 5-bit binary code u(5), and defines a scene mode parameter, scene noise intensity noise-intensity. When the current scene mode index is night mode night-mode, the scene noise intensity is encoded by entropy coding se(v).

For example, an image header example with a scene mode

Suppose the scene mode uses an Nbit integer and writes to the image header, such as the I image header or the PB image header. The code stream syntax of the image header is shown in Table 4.

Table 4 Code stream syntax of the image header containing the scene mode

The scene mode is N bits, assuming N = 5, allowing 32 common scene modes to be defined. For example, write to the I picture header I_picture_header(), PB picture header PB_picture_header(), and so on. In the above example, after the current image scene mode scene_mode, the scene mode parameter scene_freq_pattern corresponding to the current scene mode is written as a scene mode band distribution template, and the mode parameter is encoded by a 2-bit binary code. Current scene mode scene mode is night scene mode The night-scene-mode continues to encode the band parameters scene_freq_param(i), i=0...5 of the scene mode, and uses the entropy coding scheme se(v) for the band parameters of the scene mode.

For example, a stripe header example with scene mode

Taking H.264 as an example, the strip is a network packing unit and also represents the image content.

The syntax of the stripe header stream containing the scene mode is shown in Table 5.

Stream syntax for strip headers containing scene patterns

Scene—mode is N bits, assuming N = 5, allowing 32 common scene modes to be defined. For example, write to I-slice_header(), p-slice_header(), b-slicer-header().

The encoding process and decoding process of the scene mode participating in the video image will be described in detail below in conjunction with a typical video image encoder and decoder. As shown in FIG. 8, a schematic diagram of a frame-type flow reference using a typical video image encoder corresponding to transform, scaling/quantization, and entropy coding is used in the image coding method embodiment of the present invention; as shown in FIG. 9, the image of the present invention is shown in FIG. A frame-type flow reference schematic diagram of a typical video image decoder corresponding to inverse transform, scaling/inverse quantization, and entropy decoding is used in the decoding method embodiment.

In the embodiment of the image coding method of the present invention, a detailed implementation step of an image coding method based on an image mode has been given. In step 404, it is assumed that the scene mode identification value and the scene mode parameter value form defined in Table 4 are used. The scene mode parameter corresponding to the scene mode scene-mode is scene_freq_pattern, which represents a transform coefficient band distribution template of the current scene image, and the template is represented by a binary integer U(N), as shown in Table 6 below.

Table 6 Comparison table of scene mode and index value 2

Figure 10-a - Figure 10-d shows a schematic diagram of the scene freq pattern for the four scene image bands. among them,

Figure 10-a is a schematic diagram of the scene image band distribution template ( Default_ scene_freq _pattem ) in the default scene mode. It is suitable for fixed cameras. Most of the moving target objects in the scene are horizontal motion, such as moving from left to right, or from right. Move to the left;

Figure 10-b is a schematic diagram of the infrared scene mode distribution template (Infrared- scene-freq_pattern), which is suitable for camera fixation, and the shooting conditions are infrared imaging sequences;

Figure 10-c is a schematic diagram of the night scene scene distribution template (Night scene-freq_pattern) in the night scene mode, suitable for camera fixation, shooting scenes at night;

Figure 10-d is a schematic diagram of the scene image band distribution template (Moving-scene_freq_pattern) in the motion scene mode, suitable for camera fixation, the foreground object has dramatic horizontal motion (from left to right or right to left); The camera itself moves horizontally (from left to right or from right to left).

In Figure 10-a - Figure 10-d, each scene mode band distribution template has six parameter values, Fp[i], i=0,5„ where, in Figure 10, where each Fp[i] is located Indicates the frequency point of the 8x8 block. For a given i=k, the frequency point region composed of all Fp[i=k] is the kth frequency band. Therefore, Fig. 10 shows the frequency band distribution characteristics in different scenarios.

At the encoding end, as shown in Fig. 8, the scene mode scene mode and the scene mode band distribution template parameter scene_freq_pattern are used in the scaling/quantization process in image encoding. Assume that the scene mode that has been determined by the encoding end is night mode, find the corresponding band distribution template scene_freq_pattern = night scene freq pattern according to scene_mode, and press the scene_freq_pattern , value to obtain the current scene band distribution template as Figure 10-c. Shown.

It is assumed that the band distribution template scene_freq_pattern is used to discard a specific frequency band when quantizing, when Fp[i] = 0, it means that the frequency band discards i=0 ~ 5; when Fp[i] = 1 , it means that the frequency band remains, i =0 ~ 5.

The data of the transformed quantized image block is 2232 1348 991 827 499 336 190 98

1782 1186 859 732 413 317 144 83

1406 1004 705 570 355 227 102 46

1227 905 600 466 257 166 60 32

1018 687 460 350 205 102 33 7

991 666 439 309 161 71 21 4

848 494 332 232 124 56 19 2

719 438 270 182 98 52 7 1 Current scene band distribution template scene— freq_ pattern = night scene freq pattern The current band discard template is: Fp[i]={l, 1,1, 1,1,0} , ie image After the block B is quantized, the Fp[i=5] band is discarded, so the image block data that is quantized by the band control is

2232 1348 991 827 499 336 190 0

1782 1186 859 732 413 317 144 0

1406 1004 705 570 355 227 0 0

1227 905 600 466 257 0 0 0

1018 687 460 350 0 0 0 0

991 666 439 0 0 0 0 0

848 494 0 0 0 0 0 0

719 0 0 0 0 0 0 0 When encoding the scene mode, write scene_mode and scene_freq_pattern to the image header according to Table 4, and discard the template Fp[i]={l, 1 in the image header. , 1, 1, 1, 0} is written to the scene mode parameter scene_freq_param(i) , i=0...5 , in .

At the decoding end, as shown in FIG. 9, the current scene mode scene_mode is parsed from the image header shown in Table 4, and the scene mode parameter, the scene band distribution template scene_freq_pattern, and the band discard template parameter are further analyzed. Scene_freq_param(i) , i=0...5 , ie

The scene mode of the current decoded image is parsed according to the code stream scene-mode is the night mode; the corresponding band distribution template is found according to scene_mode scene_freq_pattern = night scene freq pattern;

Further parsing to obtain the band discarding template parameter scene_freq_param(i) of the current scene is

Fp[i]={l, 1,1, 1,1,0}.

The image quantized coefficient block controlled by the frequency band is BQF, and the image transform domain coefficient block obtained by inverse quantization is B. Image coding device embodiment

The image coding apparatus of this embodiment includes:

And a third module, configured to encode the image mode value, and write the encoded image mode value into the encoded code stream.

In the video coding apparatus of the embodiment of the present invention, the second module encodes the input image information by using the image mode value, that is, the scene mode characteristic, and when the scene switching occurs in the coding sequence, the method of the embodiment can be used. Adapting to the image characteristics of the transformed scene, improving the compression ratio of the image encoding. Image decoding device embodiment

The image decoding apparatus of the embodiment of the invention includes:

In the image decoding apparatus of this embodiment, the fourth module parses the image mode information existing in the code stream, and the fifth module decodes the code stream according to the image mode, thereby improving the quality of the decoded reconstructed image.

The image encoding device and the image decoding device described in the above embodiments of the present invention may constitute a system. Image mode encoding device embodiment

The image mode encoding device of the embodiment of the invention includes:

a sixth module, configured to determine an image mode value of the image to be encoded; And a seventh module, configured to encode the image mode value and write into the encoded code stream. In the image mode encoding apparatus of the embodiment of the present invention, the encoded image mode value is written into the encoded code stream by the seventh module, thereby implementing encoding of the image mode. The image mode here can be a scene mode, or other modes in an actual application. Image mode decoding device embodiment

The image mode decoding device of the embodiment of the present invention includes:

In the image mode decoding device of the embodiment, taking the scene mode as an example, the ninth module determines the scene mode corresponding to the scene mode according to the scene mode identifier value or according to the scene mode identifier value and the scene mode parameter, thereby implementing Decoding the image scene mode.

The image mode encoding device and the image mode decoding device described in the above embodiments of the present invention may constitute a system. Image coding system embodiment 1

FIG. 11 is a schematic structural diagram of Embodiment 1 of an image coding system according to the present invention. The embodiment includes an obtaining module 1, a first encoding module 2 connected to the obtaining module 1, a second encoding module 3 respectively connected to the obtaining module 1 and the first encoding module 2, and a transmission module 4 connected to the second encoding module 3. . The obtaining module 1 is configured to determine an image mode value corresponding to the image information according to the input image information of the image to be encoded; the first encoding module 2 is configured to encode the image information by using an image mode value determined by the acquiring module 1 Forming an encoded code stream; the second encoding module 3 is configured to encode the image mode value determined by the obtaining module 1 and store it in the encoded code stream; and the transmitting module 4 is configured to output the encoded code stream carrying the image mode value.

The image mode is a scene mode of an image, and the scene mode includes a scene mode identifier, or includes a scene mode identifier and a scene mode parameter.

The image encoding system of the embodiment of the present invention encodes the image information by using the image mode by the first encoding module in the video encoding, and the encoded image stream is formed by the second encoding module together with the encoded image mode, which may be encoded by the video encoding. Scene mode, controlling the encoding characteristics of video coding Parameters such as quantization parameter QP, band division mode, band parameter size, intra prediction mode, inter prediction compensation, motion estimation, and the like. When the scene switching occurs in the coding sequence, the image coding system of the present embodiment can adapt the image features of the transformed scene to improve the compression ratio of the image coding. Image coding system embodiment 2

FIG. 12 is a schematic structural diagram of Embodiment 2 of an image coding system according to the present invention. Compared with the previous embodiment, the obtaining module 1 in this embodiment may further include a first storage unit 11, a second storage unit 12, a first matching unit 13 connected to the first storage unit 11, and a first matching unit respectively. A second matching unit 14 connected to the second storage unit 12, and a third storage unit 15 connected to the first matching unit 13, the second matching unit 14, the first encoding module 2, and the second encoding module 3, respectively. The first storage unit 11 is configured to store one or more scene mode identifier values; the second storage unit 12 is configured to store a scene mode parameter value corresponding to the scene mode identifier value stored by the first storage unit 11; 13 is configured to match the scene mode identifier value corresponding to the image information in the first storage unit 11 according to the input image information of the image to be encoded; the second matching unit 14 is configured to match the scene mode according to the first matching unit 13 The identifier value matches the scene mode parameter value corresponding to the scene mode identifier value in the second storage unit 12; the third storage unit 15 is configured to store the scene pattern identifier value matched by the first matching unit 13 or store the first matching unit. 13 The matched scene mode identifier value and the scene mode parameter value matched by the second matching unit 14.

In this embodiment, the first encoding module encodes the image information by using the scene mode identifier value or the stored scene mode identifier value and the scene mode parameter value stored by the third storage unit to form an encoded code stream; The module encodes the above scene mode value and stores it in the encoded code stream; the transmission module is configured to output the encoded code stream carrying the scene mode value.

In addition, the first control unit may be further configured to determine, according to the parsed scene mode value of the image to be encoded, whether the current scene mode being used in the current decoding process has changed. When the scene mode identification value changes in the scene mode, or the corresponding scene mode parameter value changes, the current scene mode may be changed. If the current scene mode is changed, the first control unit notifies the third storage unit to update the stored corresponding scene mode identifier value, and the corresponding scene mode parameter value, that is, the first encoding module and the second encoding module use the new scenario. The mode value is image-encoded for the image to be encoded; if the current scene mode has not changed, the first control unit Notifying the third storage unit that the stored corresponding scene mode identifier value and the corresponding scene mode parameter value are not updated, that is, the first encoding module and the second encoding module continue to use the current (ie, previous time) scene mode value to perform the encoded image. Image coding.

The image encoding system of the embodiment of the present invention determines, by using the first control unit, whether the current scene mode being used in the current decoding process has changed in the video encoding, and if the current scene mode is changed, the first control unit notifies the third storage. The unit updates the corresponding scene mode identifier value stored, and the corresponding scene mode parameter value, that is, the first encoding module and the second encoding module use the new scene mode value to perform image encoding on the image to be encoded; if the current scene mode has not changed, The first control unit notifies the third storage unit not to update the stored corresponding scene mode identifier value, and the corresponding scene mode parameter value, that is, the first encoding module and the second encoding module continue to use the current (ie, previous time) scene mode. The image is encoded by the image to be encoded, so that when the scene is not switched in the encoding sequence, the scene mode used in the current encoding is not changed, the encoding time is shortened, and the encoding efficiency is improved. Image decoding system embodiment 1

FIG. 13 is a schematic structural diagram of Embodiment 1 of an image decoding system according to the present invention. This embodiment includes a parsing module 5, a decoding module 6, and an output module 7 which are sequentially connected. The parsing module 5 is configured to parse the received code stream of the image to be decoded to obtain an image mode value of the decoded image; and the decoding module 6 is configured to decode the code stream according to the image mode value obtained by the parsing module 5; The output module 7 is configured to output the decoded image of the decoding module.

The image decoding system of the embodiment of the present invention parses the scene mode information existing in the code stream by the parsing module, and the decoding module decodes the code stream according to the image scene mode information obtained by the parsing module, so that the scene mode participates in the image decoding process, thereby Improve the quality of the decoded reconstructed image. Image decoding system embodiment 2

FIG. 14 is a schematic structural diagram of Embodiment 2 of an image decoding system according to the present invention. Compared with the previous embodiment, the parsing module 5 in this embodiment may include: a first storage unit 11 and a second storage list. The element 12 is a third matching unit 51 connected to the first storage unit 11 and the second storage unit 12, respectively, and a fourth storage unit 52 connected to the third matching unit 51. The first storage unit 11 is configured to store one or more scene mode identifier values; the second storage unit 12 is configured to store a scene mode parameter value corresponding to the scene mode identifier value stored by the first storage unit 11; The 51 is configured to parse the encoded code stream of the received image to be decoded, match the scene mode identifier value in the first storage unit 11 or match the code stream in the first storage unit 11 and the second storage unit 12, respectively. Corresponding scene mode identification value and scene mode parameter value; the fourth storage unit 52 is configured to store the scene mode identifier value matched by the third matching unit 51 or store the scene mode identifier value and the scene mode parameter matched by the third matching unit 51. value.

In this embodiment, the decoding module is configured to decode the encoded code stream according to the scene mode identifier value stored by the fourth storage unit or according to the scene mode identifier and the scene mode parameter stored by the fourth storage unit.

In addition, the embodiment of the present invention may further include a second control unit, configured to determine, according to the parsed scene mode of the image to be encoded, whether the current scene mode being used in the current decoding process has changed. When the scene mode identification value changes in the scene mode, or the corresponding scene mode parameter changes, the current scene mode may be changed. If the current scene mode is changed, the second control unit notifies the fourth storage unit to update the stored corresponding scene mode identifier value, and the corresponding scene mode parameter value, that is, the code stream of the image to be decoded by the decoding module using the new scene mode. Decoding; if the current scene mode has not changed, the second control unit notifies the fourth storage unit not to update the stored corresponding scene mode identifier value, and the corresponding scene mode parameter value, that is, the decoding module continues to use the current (ie, before Once) The scene mode decodes the code stream of the decoded image.

The image decoding system of the embodiment of the present invention determines, by the second control unit, whether the current scene mode being used in the current decoding process has changed in the video image decoding. If the current scene mode changes, the second control unit notifies the fourth The storage unit updates the stored corresponding scene mode identifier value, and the corresponding scene mode parameter value, that is, the decoding module decodes the code stream of the image to be decoded using the new scene mode; if the current scene mode has not changed, the second control unit The fourth storage unit is notified not to update the stored corresponding scene mode identifier value, and the corresponding scene mode parameter value, that is, the decoding module continues to decode the code stream of the image to be decoded using the current (ie, previous time) scene mode. The image decoding system of this embodiment can change and reject according to the current scene mode. Whether to switch the scene mode in the decoding process, so that the decoding process has a certain adaptive ability to the scene switching image sequence, shortening the encoding time and improving the encoding efficiency.

Any of the above embodiments of the image coding system and the image decoding system may constitute an image coding and decoding system, and details are not described herein.

A person skilled in the art can understand that all or part of the steps of implementing the above method embodiments may be completed by using hardware related to program instructions, and the foregoing program may be stored in a computer readable storage medium, and the program is executed when executed. The method includes the steps of the foregoing method embodiments; and the foregoing storage medium includes: a medium that can store program codes, such as a ROM, a RAM, a magnetic disk, or an optical disk.

It should be noted that the above embodiments are only for explaining the technical solutions of the present invention, and are not intended to be limiting; although the present invention has been described in detail with reference to the foregoing embodiments, it will be understood by those skilled in the art that: The technical solutions described in the foregoing embodiments are modified, or some of the technical features are equivalently replaced. The modifications and substitutions do not depart from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

Rights request

An image encoding method, comprising:

2. The image encoding method according to claim 1, wherein the image mode corresponding to the image mode value comprises a scene mode of an image.

The image encoding method according to claim 1 or 2, wherein the image information includes pixel values of the image, residual values, pixel values or residual values of the image in the transform domain, and corresponding images in the image encoding. Encoding information.

The image encoding method according to claim 2, wherein the scene mode comprises a scene mode identifier; or comprises a scene mode identifier and a scene mode parameter.

The image encoding method according to claim 4, wherein the scene mode identifier comprises a scene mode identification codeword or a scene mode index.

The image encoding method according to claim 5, wherein the image encoding method further comprises: obtaining a corresponding scene mode parameter by using a scene mode identifier.

The image encoding method according to claim 4, 5 or 6, wherein the scene mode parameter is one or more of the following objects:

a feature parameter representing a feature of the scene image acquired from the image information;

Encoding information of the image information acquired in image encoding;

Control parameters present in image coding for controlling the encoder encoding process.

The image encoding method according to claim 7, wherein the feature parameters include illumination brightness of an image scene, illumination chromaticity of an image scene, coefficient band division manner of a transform domain in image coding, and coefficients of a transform domain. Band parameter values and their parameter value distribution, the way the target or background moves in the image scene.

The image encoding method according to claim 1 or 2 or 4 or 5 or 6 or 8, wherein said encoding said image information by said image mode value comprises one of the following steps or Controlling, by using the image mode value, an image block quantization mode in an encoding process; or using the image mode value to control an image block band division manner in an encoding process; or using the image mode value to control image intra prediction in an encoding process Way; or

The image inter-frame prediction mode in the encoding process is controlled by the image mode value.

The image encoding method according to claim 8, wherein the frequency band division manner comprises: a frequency band division manner in a default scene mode, a frequency band division manner in an infrared scene mode, a frequency band division manner in a night scene mode, and a motion scene. Band division mode in mode;

The typical frequency band division in the default scene mode is

— Fp[0] Fp[0] Fp[l] Fp[2] Fp[3] Fp[4] Fp[5] Fp[5]—

Fp[0] Fp[l] Fp[2] Fp[3] Fp[4] Fp[4] Fp[5] Fp[5]

Fp[l] Fp[2] Fp[3] Fp[4] Fp[5] Fp[5] Fp[5] Fp[5]

Fp[2] Fp[3] Fp[4] Fp[5] Fp[5] Fp[5] Fp[5] Fp[5]

Fp[2] Fp[4] Fp[5] Fp[5] Fp[5] Fp[5] Fp[5] Fp[5]

Fp[3] Fp[4] Fp[5] Fp[5] Fp[5] Fp[5] Fp[5] Fp[5]

Fp[4] Fp[5] Fp[5] Fp[5] Fp[5] Fp[5] Fp[5] Fp[5]_

The typical frequency band division mode in infrared scene mode is

— Fp[0] Fp[0] Fp[l] Fp[2] Fp[3] Fp[3] Fp[4] Fp[4]

Fp[0] Fp[l] Fp[2] Fp[3] Fp[4] Fp[4] Fp[5] Fp[5]

Fp[l] Fp[2] Fp[3] Fp[4] Fp[5] Fp[5] Fp[5] Fp[5]

Fp[2] Fp[4] Fp[4] Fp[4] Fp[5] Fp[5] Fp[5] Fp[5]

Fp[4] Fp[5] Fp[5] Fp[5] Fp[5] Fp[5] Fp[5] Fp[5]

Fp[5] Fp[5] Fp[5] Fp[5] Fp[5] Fp[5] Fp[5] Fp[5]

The typical frequency band division in night scene mode is

— Fp[0] Fp[0] Fp[l] Fp[2] Fp[2] Fp[2] Fp[3] Fp[5j

Fp[0] Fp[l] Fp[2] Fp[3] Fp[3] Fp[3] Fp[4] Fp[5]

Fp[l] Fp[2] Fp[3] Fp[3] Fp[4] Fp[4] Fp[5] Fp[5]

Fp[2] Fp[2] Fp[3] Fp[3] Fp[4] Fp[5] Fp[5] Fp[5]

Fp[3] Fp[3] Fp[4] Fp[4] Fp[5] Fp[5] Fp[5] Fp[5]

Fp[3] Fp[3] Fp[4] Fp[5] Fp[5] Fp[5] Fp[5] Fp[5]

Fp[4] Fp[4] Fp[5] Fp[5] Fp[5] Fp[5] Fp[5] Fp[5]

_Fp[4] Fp[5] Fp[5] Fp[5] Fp[5] Fp[5] Fp[5] Fp[5]_

The typical frequency band division in the motion scene mode is Fp[0] Fp[0] Fp[2] Fp[3] Fp[4] Fp[5] Fp[5] Fp[5]

Fp[0] Fp[l] Fp[2] Fp[3] Fp[4] Fp[5] Fp[5] Fp[5]

Fp[l] Fp[2] Fp[3] Fp[4] Fp[5] Fp[5] Fp[5] Fp[5]

Fp[2] Fp[3] Fp[4] Fp[4] Fp[5] Fp[5] Fp[5] Fp[5]

Fp[3] Fp[3] Fp[4] Fp[4] Fp[5] Fp[5] Fp[5] Fp[5]

The Fp[0], Fp[l], Fp[2], Fp[3], Fp[4], and Fp[5] are band parameters.

11. An image decoding method, comprising:

12. The image decoding method according to claim 11, wherein the image mode corresponding to the image mode value comprises a scene mode of an image.

The image decoding method according to claim 12, wherein the scene mode comprises a scene mode identifier; or comprises a scene mode identifier and a scene mode parameter.

The image decoding method according to claim 13, wherein the scene mode identifier comprises a scene mode identification codeword or a scene mode index.

The image decoding method according to claim 13 or 14, wherein the scene mode parameter is one or more of the following objects:

a feature parameter representing a scene image feature corresponding to the image information;

Encoding information of the image information acquired in image decoding;

Control parameters present in image decoding for controlling the decoder decoding process.

The image decoding method according to claim 15, wherein the feature parameters include illumination brightness of an image scene, illumination chromaticity of an image scene, coefficient band division manner of a transform domain in image coding, and coefficients of a transform domain. Band parameter values and their parameter value distribution or the way the target or background moves in the image scene.

The image decoding method according to claim 11, 12, 13, 14, or 16, wherein the decoding the code stream of the image to be decoded by using the image mode value comprises one of the following steps Or multiple: Using the image mode value to control image block inverse quantization in the decoding process; or

Controlling an image block band division manner in the decoding process by using the image mode value; or controlling image intra-frame decoding in the decoding process by using the image mode value; or

Image inter prediction compensation in the decoding process is controlled using the image mode values.

18. An image mode encoding method, comprising:

Determining an image mode value of the image to be encoded;

The image mode encoding method according to claim 18, wherein the image mode corresponding to the image mode value comprises a scene mode of the image, and the image mode value comprises a scene mode value.

The image mode encoding method according to claim 19, wherein the scene mode comprises a scene mode identifier; or comprises a scene mode identifier and a scene mode parameter.

The image mode encoding method according to claim 20, wherein the scene mode identifier comprises a scene mode identification codeword or a scene mode index.

The image mode encoding method according to claim 19, 20 or 21, wherein the scene mode comprises: an indoor scene mode, an outdoor scene mode, an infrared mode, a day mode, a night mode, a cloudy mode, and a rain Snow day mode, image gradient mode, noise mode, and perspective mode.

The image mode encoding method according to claim 19, 20 or 21, wherein the encoding the image mode value comprises:

Writing the scene mode value directly into the image encoded code stream; or

Writing the transformed value of the scene mode value into the image encoded code stream; or

The scene mode value or the transformed value of the scene mode value is entropy encoded and written into the image coded code stream.

The image mode encoding method according to claim 23, wherein the transformed value of the scene mode value is obtained by a mathematical or logical operation of the scene mode value.

The image mode encoding method according to claim 23, wherein the entropy encoding comprises a fixed length code, a variable length code or a binarized adaptive coding.

The image mode encoding method according to claim 20, wherein the pair is The scene mode values are encoded and written into the encoded code stream, including:

The scene mode value is encoded, and the encoded scene mode value is written into one or more code stream levels in the encoded code stream.

The image mode encoding method according to claim 26, wherein the code stream hierarchy comprises a code stream header structure of each level hierarchy:

Sequence header or sequence extension header;

a group header or a group extension header;

Image header or image extension header;

Strip header or strip group extension header;

Strip head or strip extension head.

28. An image mode decoding method, comprising:

An image mode is determined based on the image mode value.

The image mode decoding method according to claim 28, wherein the image mode corresponding to the image mode value comprises a scene mode of the image, and the image mode value comprises a scene mode value.

The image mode decoding method according to claim 29, wherein the scene mode comprises a scene mode identifier; or comprises a scene mode identifier and a scene mode parameter.

The image mode decoding method according to claim 30, wherein the scene mode identifier comprises a scene mode identification codeword or a scene mode index.

The image mode decoding method according to claim 30, wherein the parsing and determining the image mode value of the current decoded image from the image code stream comprises:

Parsing and determining a scene mode identification value of the currently decoded image from the image code stream;

Determining, according to the scene mode identifier value, a scene mode parameter value corresponding to the scene mode, and if the scene mode parameter value is present, decoding the scene mode parameter value;

Determining the image mode according to the image mode value includes:

Determining a scene mode according to the scene mode identification value or according to the scene mode identification value and the scene mode parameter.

33. The image mode decoding method according to claim 32, wherein the slave image The scene mode identification values that are parsed and determined in the code stream to determine the current decoded image include:

The code stream value directly read from the code stream is the parsing result value of the scene mode identifier; or the code stream value directly read from the code stream, and the parsing result of the scene pattern identifier is obtained according to the inverse transform method corresponding to the encoding end. Value; or

The code stream value directly read from the code stream is obtained by the entropy decoding method corresponding to the encoding end, and the analysis result value of the scene mode identifier is obtained.

The image mode decoding method according to claim 33, wherein the entropy decoding method comprises decoding of fixed length code decoding, variable length code decoding or binary adaptive coding method.

The image mode decoding method according to claim 32, wherein the determining, according to the scene mode identifier value, the scene mode parameter value corresponding to the scene mode comprises:

And determining, according to the scene mode identifier value, a type of the scene mode parameter corresponding to the scene mode and a value of the scene mode parameter.

The image mode decoding method according to claim 32, wherein the scene mode comprises an indoor scene mode, an outdoor scene mode, an infrared mode, a day mode, a night mode, a cloudy mode, a rainy day mode, an image Gradient mode, noise mode, or distant mode.

37. An image coding apparatus, comprising:

38. An image decoding apparatus, comprising:

39. An image mode encoding device, comprising:

a sixth module, configured to determine an image mode value of the image to be encoded; And a seventh module, configured to encode the image mode value and write into the encoded code stream.

40. An image mode decoding device, comprising:

41. An image coding system, comprising:

42. The image encoding system according to claim 41, wherein the image mode corresponding to the image mode value comprises a scene mode of an image.

43. The image coding system according to claim 42, wherein the scene mode comprises a scene mode identifier; or comprises a scene mode identifier and a scene mode parameter.

44. The image coding system according to claim 43, wherein the obtaining module comprises:

a first storage unit, configured to store one or more scene mode identifier values;

a second storage unit, configured to store a scene mode parameter value corresponding to the scene mode identifier value stored by the first storage unit;

a first matching unit, configured to match, in the first storage unit, a scene mode identifier value corresponding to the image information according to image information of an image to be encoded;

a second matching unit, configured to match, in the second storage unit, a scene mode parameter value corresponding to the scene mode identifier value according to the scene mode identifier value that is matched by the first matching unit; And storing a scene mode identifier value that is matched by the first matching unit or storing a scene mode identifier value that is matched by the first matching unit and a scene mode parameter value that is matched by the second matching unit.

45. An image decoding system, comprising: a parsing module, configured to parse a code stream of the image to be decoded, to obtain an image mode value of the image to be decoded;

a decoding module, configured to decode the code stream according to a value obtained by the parsing module; and an output module, configured to output the decoded image of the decoding module.

46. The image decoding system according to claim 45, wherein the image mode corresponding to the image mode value comprises a scene mode of an image.

47. The image decoding system according to claim 46, wherein the scene mode comprises a scene mode identifier; or comprises a scene mode identifier and a scene mode parameter.

The image decoding system according to claim 47, wherein the parsing module comprises: a first storage unit, configured to store one or more scene mode identifier values;

a third matching unit, configured to parse a code stream of the image to be decoded, match a scene mode identifier value in the first storage unit, or match a scene in the first storage unit and the second storage unit respectively Mode identification value and scene mode parameter value;

And a fourth storage unit, configured to store a scene mode identifier value that is matched by the third matching unit, or store a scene mode identifier value and a scene mode parameter value that are matched by the third matching unit.

49. An image coding and decoding system, comprising: