77 results about "Video restoration" patented technology

The invention provides an electronic image stabilization method based on characteristic coupling. The method comprises the following steps: (1) motion estimation: estimating motion parameters of a present frame and a previous frame to obtain a global motion vector and a local motion vector; (2) motion filtering: employing a Kalman filter to carry out filtering on the global motion vector, and obtaining a willful scan vector, wherein difference between the global motion vector and the willful scan vector is a vibration parameter needs to compensate; (3) motion compensation: utilizing a compensation parameter obtained by filtering carrying out frame by frame compensation; (4) video restoration: by utilizing space-time redundant information of video, carrying out restoration on the local motion vector with a motion restoration method and guiding pixel point filling. In order to solve motion carrier vibration influence on video shooting effect, the invention provides the electronic image stabilization method based on characteristic coupling. The method has the advantages of high accuracy, good anti-noise performance, strong tracking performance, good boundary consistency and the like.

The invention provides a video super-resolution processing method and device, and the method comprises the steps: obtaining each frame of image of a video, and inputting the image into a convolutionalneural network; Sequentially performing feature extraction, feature dimension reduction, nonlinear mapping and high-dimensional space mapping on the image through the convolutional neural network toobtain super-resolution features; Performing reconstruction according to the features obtained by feature extraction and the super-resolution features to obtain a super-resolution image; And finally,performing coding to form a super-resolution video code stream. The super-resolution processing of the video is realized through the convolutional neural network; Dimension reduction through features,nonlinear mapping and high-dimensional space mapping are carried out; the calculation complexity is reduced, the time complexity is reduced, the network learning difficulty is reduced and the complextexture of the output image is reserved by adopting the jump connection, so that the high reconstruction quality is realized while the real-time performance required by video processing is ensured, and the method has a very wide application prospect in the fields of video real-time transmission and compression, video restoration and the like.

The invention discloses a device eliminating interference of a television on a speech recognition device, which comprises an audio receiver connected with the audio output end of the television, an HDMI audio and video separator connected with the HDMI audio and video output end of a set top box of the television, a video output unit, a first audio output unit, an audio selector, an echo elimination unit, a speech enhancement and noise reduction unit, an audio acquisition unit, a second audio output unit, and a playback module, wherein the echo elimination unit is connected with the audio selector, the audio acquisition unit and the speech enhancement and noise reduction unit; and the audio selector is connected with the HDMI audio and video separator and the audio receiver. Technologies such as television audio signal introduction, HDMI audio and video separation, television video restoration, television audio restoration, audio selection, echo elimination, speech enhancement and noise reduction and audio acquisition are combined together, a new technical scheme in which the television and an intelligent speech device can be used at the same time is formed, and the problem that the television and the intelligent speech device can be used at the same time can be solved.

The invention relates to a video restoration method and apparatus, an electronic device and a storage medium. The method comprises the steps of obtaining a first forward optical flow and a first reverse optical flow between adjacent images in multiple continuous frames of images; performing optical flow optimization processing on the first forward optical flow and the first reverse optical flow respectively to obtain a second forward optical flow corresponding to each first forward optical flow and a second reverse optical flow corresponding to each first reverse optical flow; carrying out forward conduction optimization on the continuous multi-frame images by utilizing the second forward optical streams, and carrying out reverse conduction optimization on the continuous multi-frame imagesby utilizing the second reverse optical streams until all the images in the optimized continuous multi-frame images meet repair requirements; and according to the optimized images obtained through the forward conduction optimization and the reverse conduction optimization, obtaining the continuous restored image of the multi-frame image. According to the embodiment of the invention, the consistency of the continuous frame repair and the real-time performance of repair can be improved.

The invention discloses head-worn wireless video glasses. The head-worn wireless video glasses comprise an upper shell (1), a function button plate (2), a right door (3), a video display drive and video restoration hardware part (4), a lower shell (8), eye patches (9) and a left door (12). The head-mounted wireless video glasses are further provided with a right left-right pupil distance adjusting button (7) and a left left-right pupil distance adjusting button (10), wherein the right left-right pupil distance adjusting button (7) and the left left-right pupil distance adjusting button (10) are used for adjusting the distance between the left eye and the right eye of the glasses. By the adoption of the design, the use requirements of different crowds are greatly met, and the pupil distance between the left eye and the right eye can be adjusted. The eye patches are made of soft silica gel materials and when the eye patches are attached to the face, people feel more comfortable. Through the user-friendly design of optical lenses, a user can enjoy images with the large angle and the high definition without eyestrain. An elastic band is adopted in a head band part, the size of the elastic band is adjustable, and therefore the head-mounted wireless video glasses are suitable for anyone.

The invention discloses a monitoring video restoration method and device. The method includes the following steps that (1) an HKFS stores all video file meta information in leaf nodes, and the leftmost leaf node is obtained; (2) according to the leftmost leaf node obtained in the step (1), all the leaf nodes are traversed through deviation of other brother nodes, so that the leaf node where a deleted video file is located is found, entry information of the leaf node is obtained, a meta data area, pointed to by the entry information of the leaf node, of the video file is analyzed, and the corresponding deleted video file is restored. On the basis that a video storage structure of video monitoring is known, the monitoring video restoration method is provided. Compared with restoration methods adopted in some software in the market, the method has the advantages that restoration speed is high and restored information is complete.

The invention discloses a video encoding method, a video restoration method, a video playing system and related devices such as a server and a playing device, and relates to the technical field of computers. The video encoding method disclosed by the invention comprises the following steps: segmenting a to-be-encoded video into corresponding sub-videos according to the resolution of a plurality ofencoding devices, wherein the resolution of the sub-videos is lower than that of the to-be-encoded video; and separately inputting the sub-videos in the plurality of corresponding encoding devices for encoding. According to the video encoding method disclosed by the invention, the high-resolution to-be-encoded video is cut into a plurality of low-resolution sub-videos according to the resolutionof the plurality of encoding devices, and the sub-videos are encoded via the plurality of encoding devices, the encoding of the high-resolution video is achieved by superposing a plurality of low-resolution encoding devices, so that the improvement of the resolution of the video can be adapted by using the existing devices without upgrading or replacing the encoding devices, thereby reducing the cost and saving resources.

A method, system, and computer program for automatic restoration of digital film and video frames, by operation of a sequence of five main interactive computer program blocks, mainly separating moving objects from still objects, separating smooth texture from regions containing edges using the coherency map, and detecting and removing defects to provide restored frames using a pair of frames as reference frames for each processed frame.

A method, system, and computer program for automatic restoration of digital film and video frames, by operation of a sequence of five main interactive computer program blocks, mainly separating moving objects from still objects, separating smooth texture from regions containing edges using the coherency map, and detecting and removing defects to provide restored frames using a pair of frames as reference frames for each processed frame.

The embodiment of the invention provides a video restoration method and device and electronic equipment, and belongs to the technical field of data processing, and the method comprises the steps: responding to specific annotation information related to video restoration, and obtaining a target area set for an initial frame of a video file; in all video frames of the video file, based on an object edge constructed by a gradient value of the video frame, carrying out tracking segmentation processing on the target object by utilizing a segmentation network comprising a foreground channel and a contour channel which are distributed in parallel; executing video restoration operation on all video frames in the video file based on the result of the tracking segmentation processing; and obtaining pixel points contained in the video frame formed after the restoration operation is executed, and carrying out smoothing processing on the restored video frame according to a preset smoothing processing curve to form a restored video. Through the scheme of the invention, the video restoration efficiency is improved.

The invention discloses a video deblurring method based on a motion vector and a CNN, and the method comprises the steps: S11, carrying out the block processing of all video frames of a to-be-restoredvideo, and generating a plurality of image blocks; S12, performing fuzzy detection on a plurality of image blocks of the video frame to obtain fuzzy image blocks to be restored; S13, searching a clear image block corresponding to an adjacent frame of the to-be-restored blurred image block; S14, inputting the blurred image block to be restored and the clear image block into a CNN deblurring network; S15, the CNN deblurring network generating a deblurring result; and S16, completing the deblurring of the to-be-restored video. According to the method, the effect of efficient video blurring restoration is achieved, the method is different from the mode that deblurring is conducted on the whole video frame, and only the blurred part in the video frame is restored. In addition, the advantages of the CNN in the aspect of computer vision processing are further utilized, and a clear video restoration result is obtained.

The embodiment of the invention provides a video restoration method and device, equipment and a storage medium, and relates to the technical field of artificial intelligence, and the method specifically comprises the steps: employing a neural network model to extract the image features of a to-be-restored video frame, and predicting the image compensation parameter value of the to-be-restored video frame based on the image features. And dividing the to-be-restored video frame into a plurality of to-be-compensated regions according to the distribution characteristics of the video coding parameters corresponding to the to-be-restored video frame. And determining a region compensation parameter value of each region to be compensated according to the type of each region to be compensated, a preset image compensation strategy and the image compensation parameter value. And performing image compensation on each to-be-compensated region based on the region compensation parameter value of eachto-be-compensated region to obtain a restored video frame. Dividing the to-be-restored video frame into the plurality of to-be-compensated regions, determining the region compensation parameter valueof each to-be-compensated region based on the predicted image compensation parameter value, and carrying out targeted image compensation, so that the video restoration effect is improved.

The invention provides a rainy day video restoration method based on time domain, space domain and frequency domain joint processing. The method includes the following steps that an input video sequence is converted to YCbCr space; background modeling is conducted through a time domain median filtering method; a frame-to-frame difference sequence is calculated, and the rain scene area corresponding to each frame of image is obtained; the minimum enclosing rectangle of each raindrop in the rain scene area corresponding to each frame of image is calculated, the falling direction of each raindrop is calculated, a raindrop falling direction feature histogram is established, and the raindrop falling direction feature of each frame of image is acquired; a Y channel of each frame of image is converted into the frequency domain. In accordance with the raindrop falling direction features, an appropriate wedge-shaped direction filter is selected to conduct filtering processing on the magnitude spectra of the image, a frequency domain filtering result is inversely transformed back to the space domain, a Cb channel image and a Cr channel image of original image frames are combined, and a colored image is obtained after recombination. By means of the rainy day video restoration method based on time domain, space domain and frequency domain joint processing, the visual effect of videos can be improved obviously.

The objective of the invention is to provide a live video information playing method and device. A first P2P node device is used to determine live slice information to be loaded according to live video information requested by the first P2P node device and by combining with corresponding loading-related information, and load audio slice information and / or video slice information corresponding to the live slice information to be loaded, and then, according to the loaded audio slice information and / or video slice information, the corresponding live video restoring information is acquired to play the live video information. Compared with the prior art, the method and device of the invention can be used to realize the advantage that a user can automatically adjust the loaded video data, such as by abandoning the loading of acertain amount of non-critical video data, in different network states to make the watching of the user on the live video maintains the continuous state and will not be interrupted by data holes, the live video synchronism can be maintained, the information acquisition efficiency of the user can be improved, and the resource utilization rate of the P2P node device can be improved, so that the user experience can be improved.

The invention discloses a video restoration model training method based on a deep network and a video restoration method. The method comprises the following steps: obtaining a plurality of frames of clear images from a standard clear video; Carrying out Gaussian blur processing on each frame of clear image to obtain a blurred image corresponding to each frame of clear image; Taking a {clear imagesequence, fuzzy image sequence} pair consisting of continuous n frames of clear images and corresponding n frames of fuzzy images as a training sample, thereby obtaining a training set consisting of all training samples; Establishing a video restoration model that is connected by n-1 codec networks in turn and used for image restoration of blurred image In based on blurred image In and its previous n-1 frame blurred images In-1-I1; And training the video restoration model by using the training set so as to obtain a target video restoration model. According to the invention, the restoration efficiency and restoration effect of the blurred video can be improved.

The invention discloses a real-time video image automatic de-duplication splicing method and system, and the method comprises the steps: deploying a plurality of cameras in a target scene, dividing the target scene into a plurality of key monitoring regions, and each camera corresponding to one key monitoring region; wherein marks are arranged at four corners of each key monitoring area, an overlapping area existing between two adjacent key monitoring areas, and two marks being shared by the two key monitoring areas in the overlapping area; collecting multi-channel videos of the plurality of cameras, and storing the multi-channel videos in a centralized manner; identifying a mark of each path of video, correcting each path of video according to a mark form and pixel coordinates, and extracting an image of a key monitoring region; and analyzing and de-duplicating the multiple paths of videos according to the marks in sequence, splicing and restoring the key monitoring regions of the multiple paths of videos into one path of video, and restoring the target scene. According to the automatic de-duplication splicing of the invention, ideal spliced images can be obtained for a pluralityof static cameras or a plurality of moving cameras.

The invention provides a video restoration method, a video restoration device and a server. The method comprises the steps of determining a mask of each frame of image in a to-be-restored video; wherein the to-be-restored video comprises a to-be-removed object; covering the to-be-restored region in each corresponding frame of image in the to-be-restored video according to the mask of each frame ofimage to obtain a first video; inputting masks of each frame of image in the first video and the to-be-restored video into a first neural network model to obtain a second video; wherein the to-be-restored region in each frame of image in the second video is filled with background information of other regions except the to-be-restored region; and synthesizing the to-be-restored region in each frame of image in the second video with other regions except the to-be-restored region in each frame of image corresponding to the to-be-restored video to obtain a restored video. According to the method,the video restoration efficiency is improved.

The invention discloses a video compression method based on deep learning, and belongs to the field of video compression and deep learning. The technical problem to be solved by the invention is how to realize video compression and storage, smaller storage is realized, and the better video recovery effect is obtained. The technical scheme is adopted as follows: the method comprises the steps thatmotion estimation and motion compensation calculation are conducted through a Spynet motion estimation network composed of an optical flow network, a motion vector coding network and a motion vectordecoding network, and the better motion estimation and motion compensation effect is achieved; a residual network is used again, the residual network comprises two Resblock modules, and network training is carried out under the condition of a deeper level; and coding is completed through arithmetic entropy coding operation, the codes are stored as a Pick file, compression and storage of the videoare achieved, the smaller storage is achieved, and a better video restoration effect is obtained.

The invention discloses a video restoration method and a device, and a medium. The method comprises the following steps: obtaining a first training data set; extracting a first degradation representation of each first low-resolution video in the first training data set by using a trained feature extraction network; inputting the first degradation representation and the first low-resolution video into a pre-constructed super-resolution network to obtain a predicted high-resolution video; determining a first error between the predicted high-resolution video and the high-resolution video, and updating the super-resolution network based on the first error until the super-resolution network converges to obtain a trained super-resolution network; and when a to-be-repaired video is obtained, extracting a second degradation representation of the to-be-repaired video by using the trained feature extraction network, and inputting the second degradation representation and the to-be-repaired video into the trained super-resolution network to obtain a repaired high-resolution video. In this way, the video can be repaired according to the noise brought by different compression coding modes, and the video repairing effect is improved.

The invention discloses an ESRGAN-based video super-resolution restoration technology, device and equipment, and a storage medium. The method comprises the steps of obtaining a to-be-restored video; extracting a key frame in the to-be-restored video, and restoring an image of the key frame; grouping video frames in the to-be-restored video to generate frame groups, wherein the frame group comprises a key frame; sampling the repaired key frame in the frame group to generate a sample block; and repairing the to-be-repaired frame in the frame group through the sample block. According to the method, the data processing amount in the video restoration process is relatively reduced, so that the overall efficiency of video restoration is improved. In addition, the invention further provides a video restoration device and equipment and a storage medium, and the beneficial effects are the same as described above.

A low-latency resolution adaptive video fiber transmission codec device comprises a video input end and a video output end, wherein the video input end comprises USB equipment A, an FPGA A, an interface module A, a USB interface chip A, and a photoelectric module A; the video output end comprises a photoelectric module B, an FPGA B, an interface module B, a USB interface chip B, and USB equipmentB; the FPGA A programs to generate a video processing module, a data frame closing module A, a data de-framing module A, and an EDID processing module A; the FPGA B programs to generate a data de-framing module B, a data frame closing module B, a clock module B, a video restoration module, and an EDID processing module B; the video input end obtains video data and control timing by real-time synchronous waveform sampling of the FPGA A, and re-encodes the video data and control timing to enable synchronous transmission of video encoding and timing encoding; and the video output end performs video synchronous restoration on the received video encoding and timing encoding through the FPGA B to achieve the purpose of low delay and resolution adaptation.

The invention provides a method, system and terminal for video restoration by using a deep convolutional neural network. The method comprises the steps of carrying out the preprocessing of an originalvideo data set, and forming a training set, constructing a feature extraction network model, constructing a loss function, training the constructed feature extraction network model by jointly utilizing the training set and the loss function, and adjusting parameters of the model according to a result generated by training to obtain a final video restoration model, and repairing the video by usingthe obtained video restoration model. The invention provides a method, a system and a terminal for video restoration by using a deep convolutional neural network. The defect that a traditional methoddepends on manual definition and feature extraction is overcome, the powerful feature extraction capacity of the deep convolutional neural network is utilized, edge information between frames servesas restoration guidance, the features of video frames in the space domain and the time domain are utilized as much as possible, and meanwhile the subjective and objective quality evaluation indexes ofvideo restoration are improved.

The invention discloses a video restoration method and device, and the method comprises the steps: obtaining video information uploaded by a receiving end; decoding the video information to obtain decoded information, wherein the decoded information includes load information and data packet header information; performing recombination processing of the load information to obtain a decoded video, and performing the sequence analysis of the packet header information to obtain video frame loss information; carrying out the frame loss compensation processing of the decoded video based on the videoframe loss information to obtain a restored video. Therefore, the device does not need additional equipment, reduces software and hardware cost, and simplifies the steps of video acquisition, reducesthe distortion of the video and improves the accuracy of the restored video.

The invention provides a video restoration method, and belongs to the field of communication and information systems. The method mainly comprises the following steps of: extracting motion informationin an image frame sequence based on a mask corresponding to an original video containing a missing area; arranging a cubic sliding block for traversing and sliding in three dimension directions of theframe width, the frame height and the frame number of the original video; when the cubic sliding block slides to a mark region, randomly initializing a larger cubic sliding space; comparing the motion information with the optical flow characteristic information of the edge of the mask region, finding the most matched sample block in the sliding space range, filling and reconstructing the pixel characteristics of the mask region, continuing to traverse until all mask regions in the original video are restored, inputting the restored mask regions into a forward-backward neural network for optimization, and outputting a restored complete video. According to the method, a high-quality video restoration effect is achieved at the minimum calculation cost.

The invention discloses a scene video restoration device based on front-end object detection, comprising an object detection module, a frame sampling adjustment module and an image synthesis module. The object detection module is used for respectively detecting the object of each frame of an input video stream and sending the detected object data to the image synthesis module. The frame sampling adjusting module is used for extracting the background image from the input video stream and sending the background image and the attribute information corresponding to the background image to the image synthesizing module, and adjusting the rule of extracting background image; The image synthesizing module is used for synthesizing the target image and the background image to obtain the restored video stream. This design can ensure the accurate restoration of scene video, greatly reduce the bandwidth required to transmit images, and effectively save data storage space. In particular, this design can accurately identify and extract specific objects in the scene, in order to achieve accurate reproduction of scene details.

The invention relates to a video restoration method, a video restoration device, computer equipment and a storage medium, and belongs to the technical field of video processing. The method comprises the following steps: acquiring a to-be-restored video frame; determining the image quality damage degree of the to-be-restored video frame; if the image quality damage degree of the to-be-restored video frame is greater than or equal to a preset damage degree threshold, performing first restoration processing on the to-be-restored video frame through at least one restoration algorithm according toa serious damage restoration sequence; and if the image quality damage degree of the to-be-restored video frame is smaller than a preset damage degree threshold, performing second restoration processing on the to-be-restored video frame through at least one restoration algorithm according to the light damage restoration sequence. By adopting the method and the device, after the to-be-restored video frames are distinguished, the to-be-restored video frames can be restored through at least one restoration algorithm according to different restoration sequences. The restoration processing is relatively targeted, and the image quality of the obtained restored video frames is relatively high.

The invention discloses a fast image smog removing system based on a depth convolution neural network, which constructs an efficient single image smog removing depth convolution neural network model and a smog image training set, trains the network proposed by the invention, and finally reconstructs a clear smog-free image. By constructing deep convolution neural network model and atmospheric scattering model and training them in depth, the smog removal effect can be obtained quickly and the real-time application can be realized. The system of the present invention can be easily extended to other image or video processing fields, such as image or video deblurring, image or video restoration, and the like.

The invention discloses a video decompression method and system based on Hog features and an lgb classifier, and relates to the technical field of video decompression. The method comprises the steps of converting a monitoring video of an unmanned convenience store into a plurality of frames of images; selecting a first frame image of the original monitoring video, performing pedestrian detection by using the Hog features and the lgb classifier, and removing the pedestrian positions to form a background picture without a moving object; carrying out the pedestrian detection by using the Hog features and the lgb classifier to find out the pedestrian positions, and cutting the pedestrian positions as the pedestrian pictures to be stored and named; comparing each frame of image contained in theoriginal video with the background image to obtain a difference value, storing the residual data in a linked list, and compressing the residual data; taking out the background picture, and decoding and restoring the background of each frame of image according to the residual data in the linked list; taking out the pedestrian images in the linked list to cover each frame of background image, and decoding and restoring all frames of images; and splicing all the covered frame images into a video according to the bit rate and the code rate of the original video to complete the video restoration.

The invention relates to a video restoration method, apparatus and device, and a storage medium. Belonging to the technical field of multimedia, the embodiment of the invention provides a method capable of repairing a video. A video abstract of a video is subjected to resolution amplification processing; according to the method, the high-resolution target video abstract is obtained by extracting the high-resolution image of the video, the high-resolution image similar to each frame of image of the video is found from the target video abstract, the corresponding image in the video is restored through each frame of high-resolution image, and then the whole video can be restored by restoring each frame of image of the video. Through the method, information contained in the high-resolution image is fully utilized, and damaged details in the original image of the video are supplemented, so that the blurred video is converted into a clear video, and the quality of the video is effectively improved.

The invention relates to a video restoration method and system, and the method comprises the steps: separating a video and an audio of a to-be-restored video, and enabling the to-be-restored video tocomprise additional contents; performing frame cutting processing on the video to obtain a plurality of video frames; performing target detection on the plurality of video frames to obtain additionalcontent information; obtaining matched image information, and filling an image area where the additional content is located with the image information; and synthesizing the filled video frame and theseparated audio into a restored video. When the original video is restored, manual operation is not needed, the removal effect is good, and the efficiency is high.