104 results about "Video prediction" patented technology

A predictive coding in a video coding system, which can enhance the coding efficiency by predictively coding DC coefficients of a block to be coded using DC gradients of a plurality of previously coded neighboring blocks. The predictive coefficient is selected according to the difference between the quantized DC gradients (coefficients) of a plurality of neighboring blocks of the block to be coded, and the DC coefficient of the block to be coded is predictively coded by the selected predictive coefficient, thereby enhancing the coding efficiency.

The invention provides a method for generating and expanding the pedestrian re-identification data based on a generative network. The method includes the steps of generating a new pedestrian video frame sample by utilizing a video prediction network; generating the end-to-end pedestrian background transformation data by means of a deep generative adversarial network; expanding the breadth and therichness of the pedestrian data set by using different data generation methods; and sending the expanded data set to the feature extraction network, extracting the features and evaluating the performance through the Euclidean distance. According to the method, intra-class and inter-class data expansion of pedestrians is taken into consideration as well, more abundant samples can be generated by combining different generative networks, the expanded data set has good diversity and robustness, the problem of performance loss caused by insufficient number of samples and background interference canbe well solved, and the method has general applicability, and the expanded data set can achieve better performance and efficiency in the next step of pedestrian recognition.

Methods of encoding and decoding multi-view video by performing inter-prediction and inter-view prediction on each of pictures of multi-view video according to views are provided. A prediction encoding method of encoding a multi-view video includes determining a reference picture set; determining at least one reference list between a first reference list and a second reference list; determining one reference picture and reference block for a current block of the current picture using the determined one reference list; and performing inter-prediction or inter-view prediction for the current block using the reference block.

The invention provides a video heat prediction method based on deep belief networks. The method comprises following steps: selecting input variables according to video features; normalizing and quantifying influence factors, thus preprocessing training data; determining a single-layer restricted Boltzmann machine reconstitution dimension according to the selected input variable and feature reconstruction errors; forming the deep belief networks by multi-layer stacked restricted Boltzmann machines and a BP (Back Propagation) neural network; adjusting the deep belief networks through a global learning algorithm, thus obtaining an optimum video prediction model; placing to-be-tested video test data in the optimum video prediction model for heat prediction analysis and watching quantity prediction analysis. The invention also provides a video heat prediction system based on the deep belief networks. According to the method and the system of the invention, an online video prediction model based on the deep belief networks is provided; the deep neural network is applied in the online video prediction field; and the prediction accuracy and reliability can be improved.

Provided are a video prediction apparatus and method for a multi-format codec and a video encoding/decoding apparatus and method using the video prediction apparatus and method. The video prediction apparatus that generates a prediction block based on a motion vector and a reference frame according to a plurality of video compression formats includes an interpolation pre-processing unit and a common interpolation unit. The interpolation pre-processing unit receives video compression format information of a current block to be predicted, and extracts a block of a predetermined size to be used for interpolation from the reference frame and generates interpolation information using the motion vector. The common interpolation unit interpolates a pixel value of the extracted block or a previously interpolated pixel value in an interpolation direction according to the interpolation information to generate the prediction block.

The present invention provides an entropy coding method for coding video prediction residual coefficients, comprising the steps of: firstly, zig-zag scanning coefficients of blocks to be coded to form a sequence of (level, run) pairs; secondly, selecting a type of code table for coding a current image block to be coded according to a type of macro block; then switching and coding each (level, run) pair in the obtained sequence of (level, run) pairs with multiple tables, with the reverse zig-zag scanning order for the coding order of the pairs; at last, coding a flag of End of Block EOB with the current code table. The present invention of an entropy coding method for coding video prediction residual coefficients fully considers the context information and the rules of symbol's conditional probability distribution by designing different tables for different block types and different regions of level. The coding efficiency is improved and no impact to computational implementation complexity is involved.

The invention discloses a multi-motion stream depth convolution network model method for video prediction, which comprises the following steps: constructing a new convolution automatic coding networkframework fusing long-term and short-term memory network modules; the motion stream is proposed as the motion conversion from input frame to output frame. The method of generating multiple motion streams at the same time to learn more exquisite motion information is adopted to improve the prediction effect effectively. The base image is proposed as a pixel-level supplement to the motion flow method, which improves the robustness of the model and the overall effect of the prediction. A bilinear interpolation method is used to interact with multiple motion streams on the input frame to obtain multiple motion prediction maps, and then the motion prediction maps are linearly combined with the base image according to the weight matrix to obtain the final prediction results. By adopting the technical proposal of the invention, the time information in the video sequence can be more fully extracted and transmitted, thereby realizing longer-term, clearer and more accurate video prediction.

The invention discloses a video predictive encoding method based on neural networks, and relates to the technical field of the video compressed encoding. The method comprises the following steps: S1,inputting an encoding tree unit with a size of 64*64, and performing rough judgment on the encoding tree unit through a Bayes classifier to judge whether to adopt a SKIP mode, if yes, it is determinedthat partitioning is not performed on the current encoding tree unit, and an encoding unit size decision of the encoding tree unit is directly obtained, otherwise executing the S2; and S2, performingencoding unit blocking decision on the depth of the encoding tree unit in parallel through three neural networks to obtain a blocking result of the encoding unit; S3, obtaining an encoding unit sizedecision through the encoding unit blocking result obtained in the step S2; and S4, performing predictive encoding according to the encoding unit size decision obtained in the step S1 or S3 so as to obtain an encoding result. In the premise of ensuring the encoding performance, the encoding complexity can be greatly reduced, and the encoding efficiency is improved.

The embodiment of the invention relates to a video prediction method and a device based on face segmentation, a medium and electronic equipment, and relates to the technical field of image processing,and the method comprises the steps: carrying out prediction of a face video data set through a face segmentation model, and obtaining the true and false face segmentation label of each frame of imagein the face video data set; constructing an input data set according to every two adjacent frames of images in the face video data set and the true and false face segmentation tags corresponding to the two adjacent frames of images; adjusting the face segmentation model by using a first loss function and the input data set to obtain a target segmentation model; and segmenting a to-be-predicted face image by using the target segmentation model to obtain segmentation results, and combining the multiple segmentation results to obtain a face prediction video. According to the video prediction method, the stability of the face prediction video is improved.

The present invention provides a method of training a generator model G for video prediction so that a better, long-time video prediction effect can be obtained with less computational effort using the model. The generator model G comprises an encoder and a decoder adopting a neural network model structure, A hopping connection is adopt between that encoder and the decode, DELTA.X for generating apredicted inter-frame difference, The result of summing the predicted inter-frame difference DeltaX with the training sample is the method described in the predicted frame X<^>, includes selecting successive video frames as training samples and extracting the inter-frame difference of the training samples; 2) take that inter-frame difference as an input of an encoder in a generator model G, a formula is shown in the description wherein DELTAXi-1 is a value related to the ith inter-frame difference, Xi is the ith frame in the training sample, Xi<^> is that ith prediction frame, Xi and Xi<^> are associated with the neural network weights of the encoder and the decoder.

The present application relates to an image prediction method and apparatus based on region dynamic screening and an electronic device. According to an embodiment, an image prediction method may include dividing a current frame image into a plurality of sub-regions according to a predetermined standard; Determining an extended region corresponding to each sub-region, the extended region includingthe sub-region and being larger than the sub-region; Performing image recognition to determine an object of interest in the current frame image; Determining the sub-region in the extended region as asub-region of interest based on the extended region including at least a portion of the object of interest; Predicting the sub-region of interest using image information within the extended region toobtain a predicted sub-region of interest; And replacing a corresponding sub-region of interest in the current frame image with the predicted sub-region of interest to obtain a predicted frame image.By using the prediction method, the efficiency and speed of video prediction can be improved, and the prediction accuracy is not limited by the video size.

A video predictive coding system includes a video predictive encoding device having: an input circuit to receive pictures constituting a video sequence; an encoding circuit which conducts predictive coding of a target picture using, as reference pictures, pictures encoded and reconstructed in the past, to generate compressed picture data; a reconstruction circuit to decode the compressed picture data to reconstruct a reproduced picture; picture storage to store the reproduced picture as a reference picture for encoding of a subsequent picture; and a buffer management circuit which controls the picture storage, (prior to predictive encoding of the target picture), on the basis of buffer description information BD[k] related to reference pictures used in predictive encoding of the target picture, encodes the buffer description information BD[k] with reference to buffer description information BD[m] for a picture different from the target picture, and adds encoded data thereof to the compressed picture data.

The invention belongs to the field of 3D (three dimensional) video coding, and relates to a color video boundary-based detection and processing method for the boundary inconsistency of the depth video in the depth video coding assisted by the color video. The method comprises the concrete steps of determining the area with the boundary inconsistency between the depth video and the color video, detecting the inconsistent boundary pixels between the depth video and the color video, and processing the inconsistent boundary pixels before entropy coding. The detection and processing method can perform detection and processing on the boundary inconsistency between the color video and the depth video in the depth video prediction process based on the structural similarity between the color video and the depth video, so the rate-distortion performance of coding and the virtual viewpoint synthesis quality at a final decoding end are improved and the coding cost can be effectively reduced.

The invention provides a video coding method, which comprises the following steps: receiving a video image frame, and obtaining at least one tree coding unit of the video image frame; dividing the tree-shaped coding units according to different video prediction unit division rules so as to obtain video prediction units with different sizes of the tree-shaped coding units; performing primary selection on the predicted brightness directions of intra-frame prediction of the video prediction units of different sizes to obtain the set number of intra-frame prediction primary selection brightness directions of the video prediction units of different sizes; setting an intra-frame prediction optimal brightness direction of a video prediction unit of the current size; selecting predicted chrominance directions of intra-frame prediction of video prediction units of different sizes to obtain an intra-frame prediction optimal chrominance direction of the video prediction units of different sizes;and performing intra-frame prediction coding on the current video image frame according to the intra-frame prediction optimal brightness direction and the intra-frame prediction optimal chromaticity direction.

An embodiment of the invention provides a method for video predictive encoding. Firstly, extract a pixel block from current frame, divide the pixel block into a principal set of pixels and a complementary set of pixels; encode the principal set of pixels, output code stream of the principal set of pixels, obtain a reconstruction value of the principal set of pixels; interpolate to reconstructed pixel values, which include the reconstruction value of the principal set of pixels, take a result as predictive value of the complementary set of pixels, execute an intra-frame predictive encoding for the complementary set of pixels, output code stream of the complementary set of pixels, obtain a reconstruction value of the complementary set of pixels; and finally combine the reconstruction value of the principal set of pixels and the reconstruction value of the complementary set of pixels, and obtain a reconstruction value of the pixel block. Another embodiment of the invention also provides a device for video predictive encoding. By adopting implementation modes of the invention, compression efficiency of video encoding may be improved, implementation complexity may be lower, memory occupation may be saved, and cache hit rate may be improved. The implementation modes of the invention are suitable for highly parallel calculation, and may execute rapid encoding compression for a high-definition video source.