30 results about "Multiple description coding" patented technology

A system and method for distributing streaming content data from a server to multiple clients enables the server to handle transient surges of requests by having the clients cooperate with the server and other clients to distribute content, thereby alleviating the load on the server. The server divides the streaming content into multiple sub-streams (e.g., by using multiple description coding), and constructs multiple distribution trees with itself at the root and each client as a node in each of the trees. Each sub-stream is transmitted down a corresponding one of the distribution trees. Clients that receive sub-streams from its parent nodes in the distribution trees in turn forward the sub-streams to their child nodes in the trees.

In a digital audio broadcast system, to utilize transmission bandwidth efficiently, representations of a stereo audio signal for transmission are generated in accordance with an inventive multidescriptive coding technique. The representations, as generated, are then transmitted through multiple communication channels, respectively. The transmitted representations are received by a receiver where one or more of the representations are selected for recovery of the stereo audio signal. Because of the design of the multidescriptive coding used, the more representations are selected to recover the stereo audio signal, the higher the quality of the recovered signal.

A multi-level transform generates descriptions containing compressed data that represents source data using a description generation operation and variable support filters for compaction at each level. The initial level filters the source data and each subsequent level operates on data filtered by a prior level. The description assignment and filtering at each level may vary to create different operating points for the multi-level transform. Each operating point may have a corresponding error recovery process. In one aspect, an error recovery process encodes additional descriptions that are combined with non-erroneous description data to provide error recovery of the data in missing or damaged descriptions. In another aspect, a multi-level transform is created by combining description generation and variable support filters at the various levels.

A multi-description video encode method combining transformation and data includes applying transformation to being encoded signals and quantizing and entropy encoding the transformed signals to be decoded in terms of independent paths then to inversely transform the decoded signals to get the edge descriptions and merging the inversely transformed data to the central description, which can combines the multi-description encodes merging transformation and data with video codes. The encoding method can generate multiple MPEG code streams to one set of video sequence, and a video sequence of great distortion can be recovered from any stream, when multiple code streams are collected, a video sequence of small distortion will be recovered.

This invention discloses a multi-description coding method based on an iterative function system, which applies a tessellated segment mode to divide an iterated function system to several parts taken as descriptions to be transmitted at the coding terminal, when receiving only one path of description information, it applies a fractal extrapolation method to resume the un-received information and re-structure an iterative function system and finally repeatedly iterates the received iterative function system to get decoded images.

The invention discloses a semantic segmentation combined multi-description coding method and system. The semantic segmentation combined multi-description coding method comprises the steps that an original image and a corresponding semantic segmentation label serve as input to be transmitted to a multi-description feature generation network; the multi-description feature generation network generates two similar but not completely same descriptions by using the image context features; the generated two descriptions respectively enter a semantic segmentation coding network to be coded, and in thecoder, the descriptions are firstly up-sampled into an up-sampled image consistent with an original image in size, and then the up-sampled image and semantic segmentation mapping pass through a generation network together to generate a residual image, and the residual image and pixels of the up-sampled image are added to obtain a coarse reconstructed image; and the generated coarse reconstructedimage is transmitted to the multi-description decoding network through an information channel, and the semantic segmentation label is encoded by the standard encoder and then is transmitted to the multi-description decoding network, and a final reconstructed image is obtained through decoding. The consequence that the image reconstruction quality is not ideal due to the packet loss problem in thetransmission process is avoided, and the coding efficiency is better improved.

The invention provides a multi-description encoding and decoding method and system based on a convolutional auto-encoder. The encoding method comprises: receiving image information, carrying out convolution processing on the image information by using a convolution auto-encoder to obtain convolution characteristics, compressing the input image information into a potential space representation by the convolution auto-encoder, and then reconstructing output through the characteristics of the convolution auto-encoder, so that the output is infinitely close to the input; Dividing the obtained convolutional features into two descriptions, each description comprising two subsets, respectively encoding the two descriptions to obtain two multi-description code streams, and respectively transmitting the two multi-description code streams through different channels. On the premise that the packet loss problem is solved, the coding efficiency can be improved, and a reconstructed image with higheraccuracy can be obtained.

The invention provides an image compression method and system, a readable storage medium and computer equipment. The image compression method comprises the steps that an original image is divided intoa plurality of subsets, each subset generates a corresponding description, and meanwhile semantic segmentation labels of all the subsets are marked; on the basis of a multi-description coding method,the description corresponding to each subset and the corresponding semantic segmentation tag are reconstructed to obtain corresponding sub-compressed images; and all the sub-compressed images are synthesized to obtain a compressed image corresponding to the original image.

The invention discloses a multi-description compressed image enhancement method based on residual recursion compensation and feature fusion, belongs to the field of image quality enhancement, and solves the problem of different degrees of compression distortion of an image compressed by a multi-description coding method, especially the problem of serious structure splitting artifacts of a side decoded image. The method comprises the following steps: firstly, designing a residual recursive compensation network as a low-resolution feature extraction network of a side path and a middle path, and more effectively extracting two description decoding image features with the same content and different details by using a parameter sharing strategy; secondly, enabling the multi-description side feature up-sampling reconstruction network to adopt a network part layer parameter sharing strategy, so that the size of a network model is greatly reduced, and the generalization ability of the network is improved. Meanwhile, a multi-description middle-path feature up-sampling reconstruction network is used for performing deep feature fusion on two side-path low-resolution features and a middle-path low-resolution feature, so that efficient multi-description compressed image quality enhancement is realized, and the performance of the method is superior to that of a plurality of deep learning image enhancement methods such as ARCNN, FastARCNN and DnCNN.

The invention relates to a time domain overlap transformation multi-description coding and multi-description decoding method and system based on JND (Just Noticeable Difference). The multi-description coding method comprises the steps of obtaining image information, dividing the image information into two subsets: a first subset and a second subset, generating a first description and a second description at the same time, wherein the two descriptions comprise the first subset and the second subset at the same time; for the first description, on the basis of a multi-description method, predicting the second subset through adoption of the first subset, carrying out entropy coding according to the first subset and a prediction value of the second subset, and outputting a first multi-description code stream; and for the second description, on the basis of the multi-description method, predicting the first subset through adoption of the second subset, carrying out entropy coding according to the second subset and the prediction value of the first subset, and outputting a second multi-description code stream, wherein the multi-description method comprises the step of carrying out time domain-based overlap transformation, DCT, JND-based threshold filtering and quantization processing on the two subsets in sequence.

The invention discloses a multi-description coding method, a multi-description decoding method and a multi-description coding and decoding system based on KSVD (K singular value decomposition). The multi-description coding method comprises the steps of obtaining image information, dividing the image information into two subsets: a first subset and a second subset, and generating first descriptionand second description, wherein the two pieces of description comprise the first subset and the second subset at the same time; for the first description, on the basis of a multi-description method, predicting the second subset through adoption of the first subset, carrying out entropy coding according to the first subset and a predicted value of the second subset, and outputting a first multi-description bit stream; and for the second description, on the basis of the multi-description method, predicting the first subset through adoption of the second subset, carrying out entropy coding according to the second subset and the predicted value of the first subset, and outputting a second multi-description bit stream, wherein the multi-description method comprises the step of carrying out timedomain-based lapped transform, KSVD transform and quantization processing on the two subsets in sequence.

A method and apparatus for reducing the number of Intra-coded pictures (I-Picture) without any quality degradation. In one embodiment, the method takes advantage of characteristics of a heterogeneous network, such as Digital Subscription Line (DSL).

The invention relates to a DMB video pushing method based on multi-description coding, and belongs to the field of DMB video pushing. The method comprises the following steps: S1, decomposing a sourcevideo into a description of a low-quality video and a description of a high-quality video, and carrying out the compression and packaging; S2, pushing a video file processed in step S1 to a receivingend in a high-quality and low-quality video alternate pushing mode; and S3, verifying and decoding two sub-signals, pushed by a transmitting end, through the receiving end, and then reconstructing acorresponding video for playing. The source video is decomposed into a high-resolution description and a low-resolution description, and the two videos are pushed in a segmented and crossed manner through the DMB system, so the receiving end can play a video which is rough in image quality and can be accepted in a short time, and the video is synthesized and replaced with the corresponding low-quality segment after the high-quality video segment is received, and a long blank period of the receiving end due to waiting for receiving the high-definition video with a large memory is shortened.

The invention discloses an adaptive multimedia multipath resource distribution method in a multi-carrier communication system, which is characterized in that sub-carrier distribution is primarily determined according to channel gain status of subscriber on each sub-carrier on the basis of multi-description coding; maximum transmission rate distributed to each multi-path group by sub-carriers is computed by feature of the multi-description code, and sub-carrier distribution is regulated according to maximum transmission rate of each sub-carrier in different multi-path groups; finally power distribution is executed between sub-carriers. Using flexible decoding feature of the multi-description coding, the sub-carrier is only distributed to subscriber with higher partial channel gain in the multi-path group when distributed to one multi-path group, consequently the selection of the sub-carrier modulation order is not subjected to the limitation of subscriber with worst channel gain, so that the system spectrum efficiency is enhanced. The method supports the application on a plurality of multi-path groups, and discrete value of modulation order, thus high practicality is achieved.

The invention relates to an audio coding method and device, an audio decoding method and device, equipment, a storage medium and a product, and relates to the technical field of network communication. The method comprises the following steps: acquiring a first audio signal frame; encoding the first audio signal frame to obtain encoding parameters of the first audio signal, the encoding parameters including a whole frame type parameter and a framing type parameter; grouping the framing parameters to obtain n parameter groups; the parameter group comprises framing class parameters of at least one subframe corresponding to the parameter group, n is greater than or equal to 2, and n is a positive integer; respectively combining the whole frame type parameters with the n parameter groups to generate n description coding code streams corresponding to the n parameter groups; according to the method, single description coding equipment is transformed into multi-description coding equipment, so that the coding mode is changed from single description coding to multi-description coding, and the coding effect of the audio signal is improved while the packet loss resistance is improved.

The invention discloses an error concealment method of space-domain MDC (Multiple Description Coding). The error concealment method comprises the following steps of (1) inputting an original streaming video, and decoding the original streaming video by adopting the space-domain MDC having two descriptions; and (2) carrying out error detection on the two descriptions of the original streaming video, and judging as follows when one description is lost: a, carrying out concealment of motion information on the losing part if the losing part has time-domain correlation; b, carrying out the concealment on the losing part by adopting a space-domain interpolation method if the losing part has space-domain correlation; and c, comparing the time-domain distortion degree of the losing part with the space-domain distortion degree of the losing part: carrying out the concealment by adopting the method of step a if the time-domain distortion degree is less than the space-domain distortion degree, and carrying out the concealment by adopting the method of step b if the time-domain distortion degree is larger than or equal to the space-domain distortion degree. According to the error concealment method provided by the invention, the error concealment performance of the space-domain MDC can be enhanced.

The present invention relates to a method and apparatus to be used in designing an index assignment matrix for use in multiple description coding of an information signal. The bandwidth of the index assignment matrix is selected in dependence of transmission condition information relating to a transmission condition of a communication channel onto which a description of the information signal can be transmitted.

The invention discloses an adaptive multimedia multipath resource distribution method in a multi-carrier communication system, which is characterized in that sub-carrier distribution is primarily determined according to channel gain status of subscriber on each sub-carrier on the basis of multi-description coding; maximum transmission rate distributed to each multi-path group by sub-carriers is computed by feature of the multi-description code, and sub-carrier distribution is regulated according to maximum transmission rate of each sub-carrier in different multi-path groups; finally power distribution is executed between sub-carriers. Using flexible decoding feature of the multi-description coding, the sub-carrier is only distributed to subscriber with higher partial channel gain in the multi-path group when distributed to one multi-path group, consequently the selection of the sub-carrier modulation order is not subjected to the limitation of subscriber with worst channel gain, so that the system spectrum efficiency is enhanced. The method supports the application on a plurality of multi-path groups, and discrete value of modulation order, thus high practicality is achieved.

The invention relates to the technical field of real-time video, and provides a bit rate adaptive adjustment system applied to a cloud fusion platform, which is characterized by comprising: a media production unit, a media distribution unit, a media transcoding unit and a media consumption unit; the media The distribution unit is electrically connected to the media production unit, the media transcoding unit and the media consumption unit. The media distribution unit of the present invention can automatically adjust the transmission bandwidth according to the result of the bandwidth detection, the whole process is extremely fast, the cycle is extremely short, and the link can be quickly restored to the state of normal transmission; the media transcoding unit only converts one video source to one channel at most. The code stream of video multi-description encoding consumes less computing resources and network resources in the system, thereby reducing hardware costs; video multi-description encoding itself has a certain ability to resist packet loss, and the media consumption unit only needs to receive a sub-description to rebuild. The entire video frame can greatly improve the smoothness of video playback.

The embodiment of the invention discloses a method for lowering the redundancy rate of multi-description coding and decoding. The method comprises the following steps: acquiring a current frame masking threshold and a previous frame masking threshold; predetermining the current frame masking threshold according the current frame masking threshold and the previous frame masking threshold so as to obtain a predetermined differential signal; carrying out primary description coding on the current frame masking threshold; and carrying out secondary description coding on the predetermined differential signal. The embodiment of the invention also discloses the method and a device for reducing the redundancy rate of multi-description coding and decoding. The embodiment of the invention can lower the code rate of multiple description coding, and solves the problem of the prior art that the redundancy rate of multiple description coding is too high.

A method of wireless communication includes: an aggregator requesting a first description substream of a multimedia service from a source on a first path; receiving confirmation that the source is available; and receiving the first substream of the service. The method includes: determining whether the quality of the first sub-stream is satisfactory; and requesting to receive the multimedia service from the same source on at least one further path. A method of wireless communication by a source device comprising: receiving a request for a multimedia service from a terminal device aggregator on a first path; sending an acknowledgment that the source is available; and in response to the request, sending on the first path first substream. The method comprises receiving a request to transmit the multimedia service from the same source to the terminal device aggregator on at least one further path.

An improved (N:K) multiple description binning encoder that employs binning yet permits recovery of the input signal when fewer than K of the descriptions are available. In creating the encoder, a first choice is made of the number of descriptions that the encoder is to create and the minimum number of descriptions below which full recovery of the input signal is not possible. A second choice is made as to the number of descriptions that are to be broken up, to form descriptions that have two portions each. Once the first choice is made, appropriate quantization and binning scheme are selected by employing conventional techniques, and in response to the second choice, the chosen number of descriptions are each quantization split into coarse and fine quantization arrangements.

The present invention relates to a method and apparatus to be used in designing an index assignment matrix for use in multiple description coding of an information signal. The bandwidth of the index assignment matrix is selected in dependence of transmission condition information relating to a transmission condition of a communication channel onto which a description of the information signal can be transmitted.

A wireless communication method includes an aggregator requesting a first description substream of a multimedia service from a source over a first path, receiving confirmation that the source is available, and receiving a first substream of the service. The method includes determining if a quality of the first substream is satisfactory, and requesting to receive the multimedia service over at least one additional path from the same source. A wireless method of communication by a source device includes receiving a multimedia service request from an end device aggregator over a first path, sending an confirmation that the source is available and sending a first substream over the first path in response to the request. The method includes receiving a request to send the multimedia service over at least one additional path to the end device aggregator from the same source.

The invention relates to a time domain overlap transformation multi-description coding and multi-description decoding method and system based on JND (Just Noticeable Difference). The multi-description coding method comprises the steps of obtaining image information, dividing the image information into two subsets: a first subset and a second subset, generating a first description and a second description at the same time, wherein the two descriptions comprise the first subset and the second subset at the same time; for the first description, on the basis of a multi-description method, predicting the second subset through adoption of the first subset, carrying out entropy coding according to the first subset and a prediction value of the second subset, and outputting a first multi-description code stream; and for the second description, on the basis of the multi-description method, predicting the first subset through adoption of the second subset, carrying out entropy coding according to the second subset and the prediction value of the first subset, and outputting a second multi-description code stream, wherein the multi-description method comprises the step of carrying out time domain-based overlap transformation, DCT, JND-based threshold filtering and quantization processing on the two subsets in sequence.

A multi-level transform generates descriptions containing compressed data that represents source data using a description generation operation and variable support filters for compaction at each level. The initial level filters the source data and each subsequent level operates on data filtered by a prior level. The description assignment and filtering at each level may vary to create different operating points for the multi-level transform. Each operating point may have a corresponding error recovery process. In one aspect, an error recovery process encodes additional descriptions that are combined with non-erroneous description data to provide error recovery of the data in missing or damaged descriptions. In another aspect, a multi-level transform is created by combining description generation and variable support filters at the various levels.

The invention relates to the technical field of real-time videos, and provides a code rate adaptive adjustment system applied to a cloud fusion platform, which is characterized by comprising a media production unit, a media distribution unit, a media transcoding unit and a media consumption unit, the media distribution unit is electrically connected with the media production unit, the media transcoding unit and the media consumption unit. According to the invention, the media distribution unit can automatically adjust the sending bandwidth according to a bandwidth detection result, the whole process is very fast, the period is very short, and a link can quickly recover to a normal transmission state; the media transcoding unit is used for converting each video source into one path of video multi-description coded code stream at most, so that the consumption of computing resources and network resources in the system is low, and the hardware cost is reduced; and the video multi-description coding has certain packet loss resistance, and the media consumption unit can reconstruct the whole video frame only by receiving one sub-description, so that the video playing fluency can be greatly improved.

The present invention relates to a DMB video push method based on multiple description coding, belonging to the field of DMB video push, comprising the following steps: S1: decomposing a source video into a description of a low-quality video and a description of a high-quality video, and compressing and packaging them; S2: Push the video file processed in step S1 to the receiving end by alternately pushing high and low quality videos; S3: The receiving end checks and decodes the two sub-signals pushed by the transmitting end, and then reconstructs the corresponding video for playback . The present invention decomposes the source video into two descriptions, one with high resolution and one with low resolution, and pushes the two videos in sections through the DMB system, so that the receiving end can play a rough but acceptable video in a short time, waiting After the high-quality video segment is received, it is synthesized and replaced with the corresponding low-quality segment to shorten a long blank period at the receiving end due to waiting to receive high-definition video with a large memory.