A method and apparatus for encoding and decoding digital video data

Abstract

The present invention provides a method and device for mobile digital video encoding and decoding. The method includes: collecting video data and processing the video data to obtain mobile digital video; separating the mobile digital video to obtain a plurality of sub-frame digital images and optimizing the processing framework; encoding and decoding the sub-frame digital images based on the optimized processing framework, so as to achieve compression of digital video data during the encoding and decoding processes while ensuring high-quality video can be presented.

Description

Technical Field

[0001] The present invention relates to the technical field of digital video encoding and decoding, and particularly relates to a method and device for encoding and decoding digital video data. Background Art

[0002] At present, in daily life, video codecs are widely used. For example, in various satellite and terrestrial television broadcast systems, and on the Internet. Online video materials are usually compressed using many different codecs. In order to be able to view these materials correctly, users need to download and install a codec pack - a compiled codec component prepared for a PC.

[0003] The application scope of video technology is very wide, such as online video conferencing, online visual e-commerce, online government affairs, online shopping, online schools, telemedicine, online seminars, online exhibition halls, personal online chatting, visual consulting and other services. However, all of the above applications must be compressed, and the workload is huge.

[0004] Therefore, the present invention proposes a method and device for encoding and decoding digital video data. Summary of the Invention

[0005] The present invention provides a method for encoding and decoding digital video data, including:

[0006] Step 1, collecting video data and processing the video data to obtain mobile digital video;

[0007] Step 2, separating the mobile digital video to obtain a plurality of sub-frame digital images, and optimizing the processing framework;

[0008] Step 3, encoding and decoding the sub-frame digital images based on the optimized processing framework.

[0009] Preferably, the processing framework includes an encoder, a decoder, and an entropy parameter model module, wherein the encoder includes a main encoder and a superprior encoder, and the decoder includes a main decoder and a superprior decoder.

[0010] Preferably, in step 2, the process of separating the mobile digital video to obtain a plurality of sub-frame digital images and optimizing the processing framework further includes:

[0011] Separating the mobile digital video to obtain a plurality of sub-frames, and there is a sub-frame interval between adjacent frames, thereby obtaining a set of frame intervals of the mobile digital video;

[0012] Collect the pixel columns corresponding to each interval time point in the mobile digital video based on the set of frame intervals, obtain the relative positions of the pixel columns corresponding to each sub-frame interval, bind each sub-frame interval and the relative position of the corresponding pixel column to obtain binding information, and insert the binding information into the corresponding sub-frame interval to obtain an interval calibration block.

[0013] Preferably, in step 3, during the process of encoding and decoding the sub-frame digital image based on the optimized processing framework, it further includes:

[0014] Step 3.1, the super prior encoder and the main encoder respectively encode the interval calibration block and the image corresponding to the sub-frame to generate encoded data of a first preset length and encoded data of a second preset length;

[0015] Step 3.2, the entropy parameter model module compresses the encoded data of the first preset length and the encoded data of the second preset length;

[0016] Step 3.3, the super prior decoder and the main decoder respectively decode the compressed encoded data of the first preset length and the encoded data of the second preset length.

[0017] Preferably, in step 3, the process of encoding and decoding the sub-frame digital image based on the optimized processing framework further includes:

[0018] Perform a continuity detection on the interval calibration block;

[0019] Determine whether the adjacent frames based on the current calibration block are continuous according to the continuity detection result;

[0020] If continuous, continue to decode the adjacent frames corresponding to the next calibration block;

[0021] Otherwise, resend the encoded data of the adjacent frames of the current calibration block and give an alarm.

[0022] Preferably, in step 1, collect video data and process the video data to obtain a mobile digital video, including:

[0023] Cut the sub-video frame images in the video data into n image blocks of the same size according to a preset cutting method to obtain a number of image block data, and analyze each image block data based on a preset algorithm to obtain corresponding sub-fingerprint data;

[0024] Perform parallel processing on the n image block data after cutting the sub-video frame images, and respectively compare the fingerprints of the image block data at the same position with the image block data of the previous sub-video frame according to the unique fingerprint data;

[0025] Filter out small image block data with inconsistent fingerprint comparison from the current sub-video frame image, perform data encoding processing to obtain multiple encoded data, collect the multiple encoded data to form a data queue, and send it to the receiver through the network;

[0026] After the receiver receives the data queue, perform parallel decoding processing on the multiple encoded data therein to restore the corresponding small image block data; determine whether the restored small image block data is complete. If it is not complete, determine the vacant position, and find the small image block data corresponding to the vacant position from the previous sub-video frame image for complementation;

[0027] Stitch the complemented small image block data to generate a complete mobile digital video and display it.

[0028] Preferably, in step 2, the mobile digital video is segmented to obtain a number of sub-frame digital images, and the processing framework is optimized, including:

[0029] Perform the first optimization on the processing framework: train the processing framework based on a preset first quantizer until the processing framework meets the first preset condition for stopping training, stop training, obtain the first sub-output values of each module in the processing framework, and use the first sub-output values as the sub-initial solutions of the corresponding modules;

[0030] Process the main encoder and the pre-encoder, and adjust the first quantizer based on a preset adjustment method to obtain a second quantizer. Train the processing framework based on the second quantizer to obtain the second sub-weight values corresponding to the decoder, the entropy parameter model module, and the factorization entropy model module in the processing framework, and process the sub-initial solutions of the corresponding modules based on the second sub-weight values to obtain the first sub-optimized solutions corresponding to the decoder, the entropy parameter model module, and the factorization entropy model module;

[0031] Perform the second optimization on the processing framework: process the main decoder, the super-prior decoder, and the entropy parameter model module, and train the processing framework based on a preset third quantizer to obtain the third sub-weight values corresponding to the main encoder and the pre-encoder in the processing framework, and process the corresponding sub-initial solutions of the main encoder and the pre-encoder based on the second sub-weight values to obtain the second sub-optimized solutions corresponding to the main encoder and the pre-encoder;

[0032] Optimize the processing framework based on the first sub-optimized solution and the second sub-initial solution to obtain an optimized processing framework.

[0033] Preferably, in step 3, encode and decode the sub-frame digital images based on the optimized processing framework, including:

[0034] Based on the optimized processing framework, determine the first parameter information of the entropy parameter model, and encode the first parameter information based on the hyperprior encoder to obtain the encoded parameter information, and the hyperprior decoder decodes based on the encoded parameter information to obtain the second parameter information;

[0035] Substitute the second parameter information into the optimized processing framework to obtain the first output value and the second output value output by the entropy parameter model, and predict the bitrate of the optimized processing framework based on the first output value, the second output value and a preset algorithm to obtain the bitrate of the optimized processing framework;

[0036] Enhance the sub-frame digital image based on a preset algorithm to obtain the first sub-frame digital image, perform quantization processing on the first sub-frame digital image to obtain the second sub-frame digital image, perform image reconstruction processing on the second sub-frame digital image based on the bitrate of the optimized processing framework, output the reconstructed image, and evaluate and calculate the image optimization performance of the encoding result based on the sub-frame digital image and the reconstructed image to obtain the distortion function;

[0037] Based on the encoding result generated by optimizing the processing framework in terms of bitrate estimation, where the encoding result includes the bitrate consumed for transmitting the encoded data in the optimized processing framework and the bitrate of the second parameter information transmitted by the hyperprior decoder; construct a distortion loss function for end-to-end optimization of the image according to the distortion function and the bitrate result; perform end-to-end compression and optimization processing on the target image based on the distortion loss function.

[0038] Preferably, the hyperprior encoder is used to encode the interval calibration block to generate encoded data of a first preset length, including:

[0039] Obtain the calibration attribute and the corresponding calibration identifier of each calibration point in the interval calibration block to obtain the first symbol;

[0040] Predict and obtain the corresponding second symbol from a preset sequence according to the nature of each calibration point;

[0041] Perform a fusion process on the first symbol and the second symbol to obtain the fusion symbol corresponding to the calibration point;

[0042] Obtain the encoded data of the first preset length according to all the fusion symbols.

[0043] A device for mobile digital video encoding and decoding, including:

[0044] An acquisition unit that acquires video data and processes the video data to obtain mobile digital video;

[0045] Optimization unit, which separates the mobile digital video to obtain a number of sub-frame digital images and optimizes the processing framework;

[0046] Processing unit, which encodes and decodes the sub-frame digital images based on the optimized processing framework.

[0047] Other features and advantages of the present invention will be described in the following specification, and in part will be obvious from the specification, or will be understood by implementing the present invention. The objectives and other advantages of the present invention can be achieved and obtained by the structures specifically pointed out in the written specification, claims, and drawings.

[0048] The technical solutions of the present invention will be further described in detail below through the drawings and embodiments. Description of the Drawings

[0049] The drawings are used to provide a further understanding of the present invention, and constitute a part of the specification. Together with the embodiments of the present invention, they are used to explain the present invention and do not constitute a limitation to the present invention. In the drawings:

[0050] Figure 1 It is a flowchart of a method for encoding and decoding a mobile digital video in an embodiment of the present invention;

[0051] Figure 2 It is a flowchart of another method for encoding and decoding a mobile digital video in an embodiment of the present invention;

[0052] Figure 3 It is a structural diagram of a device for encoding and decoding a mobile digital video in an embodiment of the present invention. Detailed Embodiments

[0053] The following describes the preferred embodiments of the present invention with reference to the drawings. It should be understood that the preferred embodiments described herein are only used to illustrate and explain the present invention, and are not used to limit the present invention.

[0054] Embodiment 1

[0055] The embodiment of the present invention provides a method for encoding and decoding a mobile digital video, as Figure 1 shown, including:

[0056] Step 1, collecting video data and processing the video data to obtain a mobile digital video;

[0057] Step 2, separating the mobile digital video to obtain a number of sub-frame digital images and optimizing the processing framework;

[0058] Step 3, encoding and decoding the sub-frame digital images based on the optimized processing framework.

[0059] In this embodiment, partitioning means dividing the sub-frame digital images in the mobile digital video to obtain a number of sub-frame digital images, and inserting binding information between two adjacent sub-frame digital images to obtain an interval calibration block.

[0060] In this embodiment, the optimization processing framework means processing the modules in the processing framework respectively to obtain a first sub-optimization solution and a second sub-initial solution, and optimizing the processing framework based on the first sub-optimization solution and the second sub-initial solution, where the first sub-optimization solution and the second sub-optimization solution are the sub-initial solutions obtained after training the processing framework by the first quantizer, and are the optimization solutions corresponding to each module obtained after processing the sub-initial solutions based on the second sub-weight value and the third sub-weight value.

[0061] The beneficial effects of the above solution are as follows: By partitioning the mobile digital video and optimizing the processing framework, the redundancy in the huge data is removed, the storage space is saved, and the encoding and decoding efficiency is improved.

[0062] Embodiment 2

[0063] Based on Embodiment 1, the processing framework includes an encoder, a decoder, and an entropy parameter model module, where the encoder includes a main encoder and a superprior encoder, and the decoder includes a main decoder and a superprior decoder.

[0064] Embodiment 3

[0065] Based on Embodiment 1, during the process of the encoding and decoding method, it further includes:

[0066] Partitioning the mobile digital video to obtain a number of sub-frames, and there is a sub-frame interval between adjacent frames, and then obtaining the frame interval set of the mobile digital video;

[0067] Collecting the pixel columns corresponding to each interval time point in the mobile digital video based on the frame interval set, obtaining the relative positions of the pixel columns corresponding to each sub-frame interval, binding each sub-frame interval and the relative position of the corresponding pixel column to obtain binding information, and inserting the binding information into the corresponding sub-frame interval to obtain an interval calibration block;

[0068] In this embodiment, the interval set is a set composed of the intervals between every two adjacent sub-video frame images;

[0069] In this embodiment, the binding information means binding the relative positions of the pixels corresponding to each sub-video frame image in the mobile digital video and the sub-frame interval corresponding to the sub-video frame image to obtain binding information, and inserting the binding information into the corresponding sub-frame interval to obtain an interval calibration block, so as to facilitate the encoding and decoding of the mobile digital video;

[0070] The beneficial effects of the above solution are as follows: By segmenting the mobile digital video and inserting interval calibration blocks, the order of the mobile digital video is ensured, preventing the mobile digital video from being out of order during the encoding and decoding processes, which may affect the encoding and decoding efficiency. This guarantees the lossless reconstruction of the encoded data to be decoded and improves the encoding and decoding efficiency.

[0071] Embodiment 4

[0072] Based on Embodiment 1, as Figure 2 shown, in Step 3, during the process of encoding and decoding the sub-frame digital image based on the optimized processing framework, it further includes:

[0073] Step 3.1, the super-prior encoder and the main encoder respectively encode the interval calibration block and the image corresponding to the sub-frame to generate encoded data of a first preset length and encoded data of a second preset length;

[0074] Step 3.2, the entropy parameter model module compresses the encoded data of the first preset length and the encoded data of the second preset length;

[0075] Step 3.3, the super-prior decoder and the main decoder respectively decode the compressed encoded data of the first preset length and the encoded data of the second preset length.

[0076] In this embodiment, the first preset length and the second preset length are 4-byte length and 16-byte length respectively. Encoding the interval calibration block and the sub-video frame image into different lengths facilitates distinguishing the interval calibration block and the sub-video frame image, preventing incorrect encoding or decoding during the encoding and decoding processes, and improving the accuracy of encoding and decoding.

[0077] The beneficial effects of the above solution are as follows: The super-prior encoder and the main encoder respectively encode the interval calibration block and the image corresponding to the sub-frame, the entropy parameter model module compresses the encoded data of the first preset length and the encoded data of the second preset length; the super-prior decoder and the main decoder respectively decode the compressed encoded data of the first preset length and the encoded data of the second preset length, preventing the mobile digital video from being out of order during the encoding and decoding processes, which may affect the encoding and decoding efficiency. This guarantees the lossless reconstruction of the encoded data to be decoded and improves the encoding and decoding efficiency.

[0078] Embodiment 5

[0079] Based on Embodiment 1, in Step 3, during the process of encoding and decoding the sub-frame digital image based on the optimized processing framework, it further includes:

[0080] Performing continuity detection on the interval calibration block;

[0081] Determine whether the adjacent frames based on the current calibration block are continuous according to the continuity detection result;

[0082] If they are continuous, continue to decode the adjacent frames corresponding to the next calibration block;

[0083] Otherwise, resend the encoded data of the adjacent frames of the current calibration block and give an alarm.

[0084] The beneficial effects of the above solution are as follows: By determining whether the adjacent frames based on the current calibration block are continuous according to the continuity detection result and giving an alarm when they are not continuous, it prevents incorrect encoding or decoding during the encoding and decoding process, and improves the accuracy of encoding and decoding.

[0085] Embodiment 6

[0086] Based on Embodiment 1, in step 1, collect video data and process the video data to obtain mobile digital video, including:

[0087] Cut the sub-video frame images in the video data into n image blocks of the same size according to a preset cutting method to obtain a number of image block data, and analyze each image block data based on a preset algorithm to obtain corresponding sub-fingerprint data;

[0088] Perform parallel processing on the n image block data after cutting the sub-video frame images, and respectively compare the fingerprint of each image block data with the image block data at the same position in the previous sub-video frame image according to the unique fingerprint data;

[0089] Screen out the small image block data with inconsistent fingerprint comparison from the current sub-video frame image for data encoding processing to obtain multiple encoded data, collect the multiple encoded data to form a data queue and send it to the receiver through the network;

[0090] After the receiver receives the data queue, perform parallel decoding processing on the multiple encoded data therein to restore the corresponding small image block data; determine whether the restored small image block data is complete. If it is not complete, determine the vacant position, and find the small image block data corresponding to the vacant position from the previous sub-video frame image for complementation;

[0091] Stitch the complemented small image block data to generate a complete mobile digital video and display it.

[0092] In this embodiment, the preset algorithm is the frequency complex analysis algorithm. Analyze each image block data to obtain corresponding sub-fingerprint data, compare the sub-fingerprint data of each image block with the image block data at the same position in the previous sub-video frame image according to the unique fingerprint data, and perform data encoding processing on the small image block data with inconsistent fingerprint comparison, reducing the workload and improving the encoding and decoding efficiency.

[0093] In this embodiment, stitching means stitching the complemented small image blocks into a complete mobile digital video in the corresponding order;

[0094] The beneficial effects of the above solution are as follows: By analyzing the image block data, obtaining the sub-fingerprint data of each image block and encoding the image blocks with different sub-fingerprint data, the encoded data is more simplified, the storage space is reduced, the transmission efficiency is improved, and by establishing the correspondence between the encoded data and the image, a path and retrieval relationship for decoding the encoded data are provided for subsequent use, which facilitates the subsequent decoding of the currently obtained encoded data.

[0095] Embodiment 7

[0096] Based on Embodiment 1, in step 2, the mobile digital video is segmented to obtain a number of sub-frame digital images, and the processing framework is optimized, including:

[0097] Perform the first optimization on the processing framework: Train the processing framework based on a preset first quantizer until the processing framework meets the first preset condition for stopping training, then stop training, obtain the first sub-output values of each module in the processing framework, and use the first sub-output values as the sub-initial solutions of the corresponding modules;

[0098] Process the main encoder and the pre-encoder, and adjust the first quantizer based on a preset adjustment method to obtain a second quantizer. Train the processing framework based on the second quantizer to obtain the second sub-weight values corresponding to the decoder, the entropy parameter model module, and the factorization entropy model module in the processing framework respectively, and process the sub-initial solutions of the corresponding modules based on the second sub-weight values to obtain the first sub-optimized solutions corresponding to the decoder, the entropy parameter model module, and the factorization entropy model module;

[0099] Perform the second optimization on the processing framework: Process the main decoder, the super-prior decoder, and the entropy parameter model module, and train the processing framework based on a preset third quantizer to obtain the third sub-weight values corresponding to the main encoder and the pre-encoder in the processing framework respectively, and process the corresponding sub-initial solutions of the main encoder and the pre-encoder based on the second sub-weight values to obtain the second sub-optimized solutions corresponding to the main encoder and the pre-encoder;

[0100] Optimize the processing framework based on the first sub-optimized solutions and the second sub-initial solutions to obtain the optimized processing framework.

[0101] In this embodiment, the processing framework is an integrated encoding and decoding module composed of an encoder, a decoder, and an entropy parameter model module, which compresses, encodes, and decodes digital video data;

[0102] In this embodiment, the first quantizer performs a 4×4 integer discrete cosine transform on the digital video data and inputs the transformed digital video data into the processing framework until the processing framework meets the first preset condition for stopping training;

[0103] In this embodiment, the first preset condition is that the clarity of the digital video data encoded and decoded by the processing framework reaches within a preset range;

[0104] In this embodiment, the sub-initial solution is the first sub-output value of each module;

[0105] In this embodiment, processing the main encoder and the hyperprior encoder means freezing the main encoder and the hyperprior encoder, only running the remaining modules, obtaining the second sub-weight values corresponding to the remaining modules, and processing the sub-initial solutions of the corresponding modules based on the second sub-weight values to obtain the first sub-optimized solutions of the corresponding modules;

[0106] In this embodiment, the second quantizer performs a 2×2 integer discrete cosine transform on the digital video data and inputs the transformed digital video data into the processing framework;

[0107] In this embodiment, the third quantizer performs an 8×8 integer discrete cosine transform on the digital video data and inputs the transformed digital video data into the processing framework;

[0108] The beneficial effects of the above solution are as follows: Through the first optimization and the second optimization by the processing framework, an optimized processing framework is obtained, which simplifies the encoded data more, reduces the storage space, improves the transmission efficiency, and facilitates the subsequent decoding of the currently obtained encoded data.

[0109] Embodiment 8

[0110] Based on Embodiment 1, in step 3, encoding and decoding the sub-frame digital image based on the optimized processing framework includes:

[0111] Based on the optimized processing framework, determine the first parameter information of the entropy parameter model, and encode the first parameter information based on the hyperprior encoder to obtain the encoded parameter information, and the hyperprior decoder decodes based on the encoded parameter information to obtain the second parameter information;

[0112] Substitute the second parameter information into the optimized processing framework to obtain the first output value and the second output value output by the entropy parameter model, and perform bitrate prediction on the optimized processing framework based on the first output value, the second output value, and a preset algorithm to obtain the bitrate of the optimized processing framework;

[0113] Enhance the sub-frame digital image based on a preset algorithm to obtain the first sub-frame digital image, perform quantization processing on the first sub-frame digital image to obtain the second sub-frame digital image, perform image reconstruction processing on the second sub-frame digital image based on the bitrate of the optimized processing framework, output the reconstructed image, and evaluate and calculate the image optimization performance of the encoding result based on the sub-frame digital image and the reconstructed image to obtain the distortion function;

[0114] Based on the encoding result generated by the processing framework optimized by bitrate estimation, where the encoding result includes the bitrate consumed for transmitting the encoded data in the optimized processing framework and the bitrate of the second parameter information transmitted by the hyperprior decoder; construct a distortion loss function for end-to-end optimization of the image according to the distortion function and the bitrate result; perform end-to-end compression and optimization processing on the target image based on the distortion loss function.

[0115] In this embodiment, the first parameter information is the compression parameter of the entropy parameter model, which is used to compress the sub-frame digital image, and the compression parameter represents the compression mode of the data;

[0116] In this embodiment, the second parameter information is the parameter information obtained after encoding and decoding the first parameter information;

[0117] In this embodiment, the first output value and the second output value are two values output by the entropy parameter model after substituting the second parameter information into the optimized processing framework;

[0118] In this embodiment, the bitrate prediction is to predict the transmission speed of the optimized processing framework based on a preset bitrate prediction algorithm and the first output value and the second output value to obtain the data transmission speed of the optimized processing framework, that is, the bitrate;

[0119] In this embodiment, the quantization processing is to divide the first sub-frame digital image into several intervals and replace each interval with the average pixel value of the interval, so as to simplify the first sub-frame digital image and obtain the second sub-frame digital image;

[0120] In this embodiment, the image reconstruction processing is to convert the second sub-frame digital image from digital information to image information, and evaluate and calculate the image optimization performance of the encoding result to obtain the distortion function;

[0121] In this embodiment, the distortion function is the distortion value between the image output by the optimized processing framework and the input image;

[0122] The beneficial effects of the above solution are as follows: Based on the optimized processing framework, the first parameter information of the entropy parameter model is determined, and the first parameter information is decoded to obtain the second parameter information. Then, based on the second parameter information, the bit rate of the optimized processing framework is predicted to obtain the distortion function. Finally, based on the distortion function, end-to-end compression and optimization processing are performed on the target image, improving the lossless restoration degree of the image during the encoding and decoding processes.

[0123] Example 9

[0124] Based on Example 7, during the process of image reconstruction of the second sub-frame digital image based on the bit rate of the optimized processing framework, it includes:

[0125] According to the following formula, calculate the first loss value of image reconstruction:

[0126]

[0127] where D 1 is the first loss value of image reconstruction, i is the number of pixels of the second sub-frame digital image, j is a preset weight value, is the predicted two-dimensional coordinate corresponding to each pixel, is the marked two-dimensional coordinate corresponding to each pixel;

[0128] Through the formula the first loss function of image reconstruction can be obtained. For example, when i = 1000 and j = 1, then D 1 = 1.998, and according to the following formula, calculate the second loss value of image reconstruction:

[0129]

[0130] where, D 2 is the second loss value of image reconstruction, is the predicted reconstruction error corresponding to each pixel;

[0131] Through the formula the second loss value of image reconstruction can be obtained. For example then D 2 = 0.5. Based on the second loss value of image reconstruction, determine whether the image reconstruction meets the standard;

[0132] When D 2 is less than the preset threshold, it indicates that the reconstruction of the second sub-frame digital image is qualified and no processing is performed;

[0133] When D 2 is not less than the preset threshold, it indicates that the reconstruction of the second sub-frame digital image does not meet the standard, and the image is reconstructed twice until the image reconstruction meets the standard.

[0134] The beneficial effects of the above solution are as follows: Based on the distortion function, it is determined whether the image reconstruction meets the standard. When the image reconstruction does not meet the standard, secondary reconstruction is performed until the image reconstruction meets the standard, improving the lossless restoration degree of the image during the encoding and decoding processes.

[0135] Example 10

[0136] Based on Example 4, the superprior encoder is used to encode the interval calibration block to generate encoded data of a first preset length, including:

[0137] Obtain the calibration attribute and corresponding calibration identifier of each calibration point in the interval calibration block to obtain a first symbol;

[0138] According to the nature of each calibration point, predict and obtain a corresponding second symbol from a preset sequence;

[0139] Fuse the first symbol and the second symbol to obtain a fused symbol corresponding to the calibration point;

[0140] According to all the fused symbols, obtain the encoded data of the first preset length.

[0141] The beneficial effects of the above solution are as follows: It reduces the workload of encoding and decoding, improves the encoding and decoding efficiency, and ensures the lossless restoration of the encoded data to be decoded.

[0142] Example 11

[0143] An embodiment of the present invention provides a device for mobile digital video encoding and decoding, as Figure 3 shown, including:

[0144] An acquisition unit that acquires video data and processes the video data to obtain mobile digital video;

[0145] An optimization unit that divides the mobile digital video to obtain a number of sub-frame digital images and optimizes the processing framework;

[0146] A processing unit that encodes and decodes the sub-frame digital images based on the optimized processing framework.

[0147] The beneficial effects of the above solution have been described in Example 1.

[0148] Obviously, those skilled in the art can make various changes and modifications to the present invention without departing from the spirit and scope of the present invention. Thus, if these modifications and variations of the present invention fall within the scope of the claims of the present invention and their equivalent technologies, the present invention is also intended to include these modifications and variations.

Claims

1. A method for mobile digital video encoding and decoding, characterized in that, it includes: Step 1, collect video data and process the video data to obtain mobile digital video; Step 2, separate the mobile digital video to obtain a number of sub-frame digital images, and optimize the processing framework; Step 3, encode and decode the sub-frame digital images based on the optimized processing framework; Step 2, when separating the mobile digital video to obtain a number of sub-frame digital images and optimizing the processing framework, it further includes: Separate the mobile digital video to obtain a number of sub-frames, and there is a sub-frame interval between adjacent frames, and then obtain the frame interval set of the mobile digital video; Collect the pixel columns corresponding to each interval time point in the mobile digital video based on the frame interval set, obtain the relative positions of the pixel columns corresponding to each sub-frame interval, bind each sub-frame interval and the relative positions of the corresponding pixel columns to obtain binding information, and insert the binding information into the corresponding sub-frame interval to obtain an interval calibration block; Step 2, separate the mobile digital video to obtain a number of sub-frame digital images and optimize the processing framework, including: Perform the first optimization on the processing framework: train the processing framework based on a preset first quantizer until the processing framework meets the first preset condition for stopping training, stop training, obtain the first sub-output values of each module in the processing framework, and use the first sub-output values as the sub-initial solutions of the corresponding modules; Process the main encoder and the pre-encoder, and adjust the first quantizer based on a preset adjustment method to obtain a second quantizer. Train the processing framework based on the second quantizer to obtain the second sub-weight values corresponding to the decoder, the entropy parameter model module, and the factorization entropy model module in the processing framework respectively, and process the sub-initial solutions of the corresponding modules based on the second sub-weight values to obtain the first sub-optimized solutions corresponding to the decoder, the entropy parameter model module, and the factorization entropy model module; Perform the second optimization on the processing framework: process the main decoder, the super-prior decoder, and the entropy parameter model module, and train the processing framework based on a preset third quantizer to obtain the third sub-weight values corresponding to the main encoder and the pre-encoder in the processing framework respectively, and process the corresponding sub-initial solutions of the main encoder and the pre-encoder based on the third sub-weight values to obtain the second sub-optimized solutions corresponding to the main encoder and the pre-encoder; Optimize the processing framework based on the first sub-optimized solution and the second sub-initial solution to obtain the optimized processing framework; The processing framework includes an encoder, a decoder, and an entropy parameter model module, where the encoder includes a main encoder and a super-prior encoder, and the decoder includes a main decoder and a super-prior decoder.

2. The method for mobile digital video encoding and decoding according to claim 1, characterized in that: In the process of encoding and decoding the sub-frame digital images based on the optimized processing framework in Step 3, it further includes: Step 3.1, the super-prior encoder and the main encoder respectively encode the interval calibration block and the image corresponding to the sub-frame to generate encoded data of a first preset length and encoded data of a second preset length; Step 3.2, the entropy parameter model module performs data compression on the encoded data of the first preset length and the encoded data of the second preset length; Step 3.3, the hyperprior decoder and the main decoder respectively decode the compressed encoded data of the first preset length and the encoded data of the second preset length.

3. A method for mobile digital video encoding and decoding according to claim 1, characterized in that in step 3, during the process of encoding and decoding the sub-frame digital image based on the optimized processing framework, it further includes: performing continuity detection on the interval calibration block; judging whether the adjacent frames based on the current calibration block are continuous according to the continuity detection result; if continuous, continue to decode the adjacent frames corresponding to the next calibration block; otherwise, resend the encoded data of the adjacent frames of the current calibration block and give an alarm.

4. A method for mobile digital video encoding and decoding according to claim 1, in step 1, collecting video data and processing the video data to obtain a mobile digital video, including: cutting the sub-video frame images in the video data into n image blocks of the same size according to a preset cutting method to obtain a number of image block data, and analyzing each image block data based on a preset algorithm to obtain corresponding sub-fingerprint data; performing parallel processing on the n image block data after cutting the sub-video frame images, and respectively comparing the fingerprint with the image block data at the same position in the previous sub-video frame image according to the unique fingerprint data; screening out the small image block data with inconsistent fingerprint comparison from the current sub-video frame image for data encoding processing to obtain a plurality of encoded data, collecting the plurality of encoded data to form a data queue and sending it to the receiving party through the network; after the receiving party receives the data queue, performing parallel decoding processing on the plurality of encoded data therein to restore the corresponding small image block data; judging whether the restored small image block data is complete, if not, determining the vacant position, and finding out the small image block data corresponding to the vacant position from the previous sub-video frame image for complementation; stitching the complemented small image block data to generate a complete mobile digital video and displaying it.

5. A method for mobile digital video encoding and decoding according to claim 1, in step 3, encoding and decoding the sub-frame digital image based on the optimized processing framework, including: based on the optimized processing framework, determining the first parameter information of the entropy parameter model, and encoding the first parameter information based on the hyperprior encoder to obtain the encoded parameter information, and the hyperprior decoder decodes based on the encoded parameter information to obtain the second parameter information; substituting the second parameter information into the optimized processing framework to obtain the first output value and the second output value output by the entropy parameter model, and predicting the coding rate of the optimized processing framework based on the first output value, the second output value and a preset algorithm to obtain the coding rate of the optimized processing framework; Enhance the sub-frame digital image based on a preset algorithm to obtain the first sub-frame digital image, perform quantization processing on the first sub-frame digital image to obtain the second sub-frame digital image, perform image reconstruction processing on the second sub-frame digital image based on the bitrate of the optimized processing framework, output the reconstructed image, and perform image optimization performance evaluation calculation on the encoding result based on the sub-frame digital image and the reconstructed image to obtain a distortion function; Based on the bitrate result generated by the processing framework optimized by bitrate estimation, where the bitrate result includes the bitrate consumed for transmitting encoded data in the optimized processing framework and the bitrate of the second parameter information transmitted by the hyperprior decoder; construct a distortion loss function for end-to-end optimization of the image according to the distortion function and the bitrate result; perform end-to-end compression and optimization processing on the target image based on the distortion loss function.

6. A method for mobile digital video encoding and decoding according to claim 2, characterized in that, The hyperprior encoder is used to encode the interval calibration block to generate encoded data of a first preset length, including: Obtain the calibration attribute and the corresponding calibration identifier of each calibration point in the interval calibration block to obtain a first symbol; Predict and obtain the corresponding second symbol from a preset sequence according to the nature of each calibration point; Perform fusion processing on the first symbol and the second symbol to obtain a fusion symbol corresponding to the calibration point; Obtain encoded data of a first preset length according to all the fusion symbols.

7. A mobile digital video encoding and decoding device, the device executes the method for mobile digital video encoding and decoding according to claim 1, characterized in that, including: An acquisition unit that acquires video data and processes the video data to obtain a mobile digital video; An optimization unit that separates the mobile digital video to obtain a plurality of sub-frame digital images and optimizes the processing framework; A processing unit that encodes and decodes the sub-frame digital image based on the optimized processing framework.

Images