161 results about "Video Library" patented technology

Systems, methods and articles of manufacture for modeling a joint language-visual space. A textual query to be evaluated relative to a video library is received from a requesting entity. The video library contains a plurality of instances of video content. One or more instances of video content from the video library that correspond to the textual query are determined, by analyzing the textual query using a data model that includes a soft-attention neural network module that is jointly trained with a language Long Short-term Memory (LSTM) neural network module and a video LSTM neural network module. At least an indication of the one or more instances of video content is returned to the requesting entity.

The embodiment of the invention discloses a video clip detecting method and device, and belongs to the field of image processing. The video clip detecting method comprises the steps that a currently-input video file or real-time streaming video is used as an object to be detected, whether video in a video library appears in the current video stream or not is detected and recognized through the video analysis and retrieval method, the accurate position, in the stream to be detected, of a video clip is located and a detection result is output. Image features extracted through the video clip detecting method and device integrate image global information and image local information, the matching error rate is reduced while the high recall rate is guaranteed, robustness is kept for image content changes, therefore, accurate locating is conducted on video clips in real-time or off-line television programs or video programs and accurate detection and recognition of the video clips are kept on the conditions that frames of video streams are lost, mosaic blocks appear in the video streams, video resolution ratios and video frame rates are different and video contents change.

The invention discloses a video detection method. The method comprises the steps of obtaining a to-be-detected target video; performing feature extraction on each key frame in the target video to generate a target video fingerprint, the target video fingerprint comprising the image features of each key frame in the target video; according to the target video fingerprint and a reference video fingerprint corresponding to a reference video in a reference video library, determining the distance from each key frame in the target video to the reference video, wherein the reference video fingerprint comprises the image characteristics of each key frame in the reference video; and determining the similarity between the target video and the reference video according to the distance between each key frame in the target video and the reference video. Due to the fact that the local space features and the local frame arrangement time sequence information of all the key frames are reserved, the spatial repetitive inspection is more accurate, and the long and short video mutual inspection and the similar video time position positioning are supported. The invention further discloses a corresponding device, an apparatus and a medium.

The invention relates to a method for retrieving similar video clips based on sports competition videos. The method comprises the following steps of: preprocessing all videos in a video library by using an off-line part to form a video feature library; performing video structured analysis and feature extraction on the queried clips by using an on-line part; and performing two-round vari-size grained retrieval in the video feature library, wherein a candidate video set is determined in the first-round coarse-grained retrieval process, and the accurate positions of the similar clips in a video object are determined in the second-round fine-grained retrieval process. The query process is accelerated by organizing all video features through a K-dimension (K-D) tree; and the influence of vision, a time sequence and an interference factor on similarities is comprehensively considered, and the similarities of the similar chips are calculated and sequenced.

A display interface device and method for an individual personal computer, workstation, or embedded system comprising one or more display interfaces; a video library; and driving any one of a plurality of video displays of a plurality of types connected to the one or more display interfaces from output of the video library. OpenGL is a preferred API. Driving of hybrid displays as well as stroke displays using formats designed for raster displays (such as VAPS) is also provided for. Real-time switching between displays is also provided for.

The invention relates to the field of video information, and discloses a video clustering method, an ordering method, a video searching method and corresponding devices. The video clustering method comprises a clustering step used for clustering multiple pixels of which the similarity is higher than a predetermined threshold value together by aiming at each frame in a video, so as to obtain super pixels capable of imitating the minimum entity of human vision, wherein the similarity is calculated according to the colors, the positions and / or the motion characteristics of the pixels. Through conducting time-space domain self-correlation and cross-correlation analysis on pixel areas in the video and ordering the pixel areas in the video on the basis of the super pixels capable of imitating the minimum entity of the human vision, the invention has the advantage that videos can be accurately searched in a real-time manner even in a massive video database.

The present disclosure relates to a video processing method and apparatus, an electronic device and a storage medium. The method includes: determining a preselected video associated with a query textparagraph among a plurality of videos according to paragraph information of a query text paragraph and video information of a plurality of videos in a video library; the target video in the preselected video is determined according to the video frame information of the preselected video and the sentence information of the query text paragraph. According to the video processing method of the embodiment of the present disclosure, the video can be retrieved through the correlation between the video and the query text paragraph, the target video can be accurately searched, the redundancy of the query result can be avoided, and the query text paragraph in the natural language form can be processed without being limited by the inherent content of the content label.

The embodiment of the invention discloses a video duplicate checking method and device, and the method comprises the steps: constructing a multi-modal feature vector of a to-be-processed video; carrying out neighbor retrieval in a video library based on the multi-modal feature vector, screening out candidate videos similar to the to-be-processed video, and obtianing a candidate video set; calculating the similarity between each candidate video and the to-be-processed video to obtain a similarity result; and determining whether the to-be-processed video passes duplicate checking detection or not according to the similarity result. According to the scheme, the video duplicate checking efficiency can be improved while the video duplicate checking accuracy is ensured.

The invention discloses an RFID (Radio Frequency Identification) reader and camera integrated machine. The integrated machine comprises an ultrahigh-frequency RFID antenna, an ultrahigh-frequency RFID read unit, a camera, an image processing unit, a public bus, a processor, a memory, a communication interface and a communication link, wherein the ultrahigh-frequency RFID antenna is connected with the ultrahigh-frequency RFID read unit, the camera is connected with the image processing unit, and the communication interface is connected with the communication link; and the ultrahigh-frequency RFID antenna, the image processing unit, the processor, the memory and the communication interface are all connected with the public bus. Compared with the prior art, the integrated machine has the advantages of effectively improving the working efficiency and the accuracy and effectively improving the processing speed of user's multi-target and long-time photography or searching a desired target photo or target video frame from a video library in the field of automotive electronic identification. Meanwhile, the integrated machine is easily applied in other fields of internet of things, and is suitable for large-scale promotion.

The invention discloses a group behavior analysis method based on multi-feature fusion, and belongs to the field of machine vision and intelligent information processing. On one hand, the method researches the group behavior characteristics of a group level, one series of feature descriptors which represent local movement information can be extracted, wherein the feature descriptors comprise gregariousness, stability, consistency and conflict; on the other hand, the method imports a new multidimensional optical flow histogram feature to represent the global movement information, and a multilayer dictionary learning method to carry out further optimization; and finally, through a local movement feature descriptor and a global movement feature descriptor are fused to form a feature set whichcan comprehensively describe group behaviors, and the method can be applied to aspects including group behavior analysis, behavior recognition and the like. The rationality and the effectiveness of the method can be proved through experiments on a real video library.

The invention discloses a method and system for playing a pop-up screen in a local video. The method comprises the steps that a header file among local video files is intercepted, the header file is compressed, and a corresponding header file MD5 value is generated; the local video files with the header file MD5 value are identified and collected to form a video library; all users who watch any local video file in the video library locally are collected, and a same-block user set is established; pop-up screen information sent by all the users in the same-block user set to the local video files through a pop-up screen sending area is collected; the header file MD5 value of the local video files and the pop-up screen information sent by the users who have watched the local vide are bound and uploaded; the bound header file MD5 value and pop-up screen information are analyzed and stored into a pop-up screen information sending module; the stored pop-up screen information is sent into all the local video files in the video library to be displayed.

The invention discloses a video label generation method and device and electronic equipment, and relates to the technical field of image processing and the technical field of natural language processing. The specific implementation scheme comprises the steps of obtaining a to-be-processed target video; extracting a key video frame from the target video; obtaining candidate reference videos from areference video library, wherein the reference video library comprises a plurality of reference videos and text information corresponding to each reference video, and the similarity between at least one video frame in the candidate reference videos and the key video frame is larger than a first threshold value; and determining a label of the target video according to the text information corresponding to the candidate reference video. Therefore, through the artificial intelligence, especially the video label generation method based on deep learning, fine-grained identification of the target video does not need to be carried out by using an image classification model, so that the accuracy and recall rate of video fine-grained identification are improved, the reference video in the referencevideo library can be expanded at any time, and the expandability is good.

The invention provides a system and method for prevention and control learning of hospital infection by medical staff, and belongs to the technical field of medical treatment. The system comprises a full-time personnel function and post module. The system comprises a learning curriculum module, a practice and evaluation module and a score and skill post authentication module, and a hospital infection prevention and control knowledge base in the learning curriculum module comprises a hospital infection professional curriculum library, a hospital infection standard operation video library, a hospital infection capability assessment question library and a hospital infection related laws and regulations knowledge question library; Configuring corresponding ability assessment question banks according to different professional medical staff, randomly generating examination questions and pushing the examination questions, and assessing the hospital infection ability of the medical staff; Self-learning and self-evaluation are carried out by medical staff at any time and any place in a specified task period; And an electronic certificate is issued to the medical staff who completes the score learning and examination. An online learning platform is constructed, learning courses are pushed out for individuals, and medical staff can learn the learning courses anytime and anywhere.

The invention relates to a video matching method and device, computer equipment and a storage medium. The method comprises the following steps: inputting an acquired video source into a feature extraction network for feature extraction to obtain a feature map and a feature vector corresponding to each video frame in the video source, wherein the video source comprises at least one video frame, andeach video frame comprises a to-be-detected object; inputting the feature map corresponding to each video frame into a similarity extraction network for similarity calculation to obtain the similarity between every two adjacent video frames in the video source; performing fusion processing on the feature vectors corresponding to the video frames based on the similarity between every two adjacentvideo frames to obtain fusion features corresponding to the video source; determining a target video sequence matched with the video source in a video library according to the fusion feature corresponding to the video source and the preset video library, wherein the target video sequence comprises at least two target video sources. By adopting the method, the accuracy of a video matching result can be improved.

The invention belongs to the 3D video processing field of the new generation information technology based on the augmented reality, to be specific, relates to a 360-degree multi-viewpoint 3D holographic video acquisition method, and a realization method. A plurality of video cameras are used for the full-color, high resolution, and all-dimensional image acquisition, pixel operation and pixel aggregation, and therefore the 3D holographic video acquisition way and the processing procedure can be simplified, the consumption of a lot of time and dedicated devices can be saved, the all-dimensional and full-color 3D holographic videos can be acquired, and the spatial 3D visual sense of the user can be greatly enhanced. The video cameras adopt the MAC architectures, and can share the channel in a way of combining the statistical multiplexing and the assembly line transmission, and therefore the rapidness and the effectiveness of the data transmission can be improved, and the resource waste condition and the channel conflict condition can be prevented. The video database and the user terminal adopt the single-instruction multi-channel transmission way, and then the wireless resource transmission unified allocation can be realized, and therefore the data transmission speed can be accelerated, and the 3D holographic video processing and display speed of the user terminal can be accelerated.

The embodiment of the invention discloses a video clip retrieval method and device, and relates to the field of video data search. The method comprises the steps of performing video frame sampling ona target video and a to-be-retrieved video in a video library, and performing feature extraction on sampled sampling frames; matching the target video with the to-be-retrieved video according to the characteristics of the sampling frame, and determining candidate to-be-retrieved videos matched with the target video; determining at least one candidate video clip from the candidate videos to be retrieved, and calculating the matching degree of the target video and each candidate video clip based on the characteristics of each sampling frame; and determining a video clip matched with the target video in the to-be-retrieved video according to the matching degree of the target video and each candidate video clip. The method improves the accuracy of video retrieval.

The invention discloses a multi-person behavior detection method and system based on deep learning fusion of various interaction information. The method comprises the following steps: building a videolibrary with a label as a sample set to train a behavior detection network, processing a to-be-detected video through the trained network, and realizing the detection of an object behavior in a region according to a final output vector. According to the method, the complexity of human behaviors is fully considered, the interaction relationship between the human behaviors and other people, objectsand long-term memory information is integrated while the motion of the human is considered, and the video behavior detection precision is effectively improved.

The embodiment of the invention provides a video recommendation method and device and electronic equipment, and is applied to the technical field of the Internet. The method comprises the following steps: acquiring a plurality of target labels; for each target label, obtaining a popularity value of the target label and a first interest value corresponding to the target label; based on the total number of users of the target client and the target number of users accessing the video with the target label in the target client, determining an importance degree value of the target label for distinguishing user interests, and calculating a label weight of the target label by utilizing the popularity value of the target label and the corresponding first interest value as well as the importance degree value of the target label for distinguishing the user interests; selecting a target label for video recommendation based on the label weight of each target label; and based on the selected targetlabel, selecting a recommended video for the target user from a video library, and recommending the recommended video to the target user. According to the scheme, the problem of low accuracy of videorecommendation in the prior art can be solved.

The invention relates to the technical field of image processing technology and pattern recognition, and discloses a method for counting video playing times. The method comprises the following steps of: S1, defining a video image sequence F={f1,f2,f3,ellipsis,fN}, and extracting the feature vectors of key frames to obtain a feature vector set adv={V1,V2,ellipsis,VM} and a feature database adv_db={adv1,adv2,ellipsis,advL}; and S2, detecting a certain quantity of videos, and counting the video playing times according to the feature database. In the method, a video is presented by adopting a group of key features, and a function of automatically extracting key features from the video is realized by using a method for combining a multi-scale grid with a haar feature, so that image content matching of real-time videos and template videos is realized; and moreover, the playing times of video advertisements are counted by using a key feature frequency histogram accumulated integration method, and key feature vectors are short and are rapid for matching, so that the method is very suitable for searching and comparing large-scale video libraries, and is suitable for counting video playing times.

The invention provides a video retrieval method, device and equipment and a storage medium. The method comprises the following steps: acquiring a comparison video clip from a video library according to the time length of a to-be-tested video; determining the similarity between the to-be-tested video and the comparison video clip through a space-time neural network, wherein a space-time convolutionlayer of the space-time neural network is set to carry out two-dimensional convolution and time dimension information processing; and traversing the video library, and outputting a retrieval result according to the similarity. In the process of determining the similarity between the to-be-tested video and the comparison video clip, the space-time neural network is used for determining the similarity between the to-be-tested video and the comparison video clip; a space-time convolution layer of the space-time neural network firstly performs two-dimensional convolution, processes spatial features of a to-be-detected video and a comparison video clip, then processes time dimension information of the to-be-detected video and the comparison video clip, and realizes low-complexity space-time feature extraction in a mode of successively processing the spatial features and the time dimension information.

The invention discloses a video processing method and device, a storage medium and an electronic device. The method comprises steps of acquiring a to-be-processed first video wherein the first video comprises a first video frame with a first definition; based on a first video relation chain matching a first video frame, searching a second video frame matching the first video frame from a video library, wherein the similarity between the content of the second video frame and the content of the first video frame is greater than a first threshold, the definition of the second video frame is a second definition, the second definition is greater than the first definition, and at least one video frame of which the similarity with the first video frame is greater than a second threshold exists ineach video in the first video relation chain; and replacing the first video frame in the first video with the second video frame. According to the invention, the technical problem of low efficiency of improving the definition of the video in related technologies is solved.

The embodiment of the invention provides a video recommendation method and device and electronic equipment, and the method comprises the steps: determining at least one target publisher according to the feature information of a plurality of video publishers, the video features of a plurality of videos and a user watching behavior log, and generating a target publisher set; and combining the videosin the video library according to the video pushing relationship, obtaining the click rate of the second video in each video combination according to the user watching behavior log and the video pushing relationship to generate a click rate list, and for each target publisher in the target publisher set, when the first target video is played, obtaining a video recommendation list corresponding tothe first target video, sorting the to-be-recommended videos in the video recommendation list according to the click rate list, and screening out a first preset number of second target videos for recommendation. According to the invention, the experience effect of the user in the publisher recommendation scene can be improved, recommendation can be performed according to the video playing condition, and the watching experience of the user is ensured.

The invention provides a video recommendation method and system based on a deep neural network, and a storage medium. According to the method and system, the watching video data of a user are obtained; a positive sample data set and a negative sample data set are generated according to the watching video data of the user and the video data of a video library; autoencoder model training based on adeep neural network is performed with the positive sample data set and the negative sample data set, so that a training network model can be obtained; the video watching data of each user and the video data of the video library are inputted into the training network model, so that a predicted video recommendation list can be obtained; and video recommendation is performed on each user according tothe predicted recommended video list. According to the method provided by the invention, personalized recommendation under a network television condition is realized, recommendation accuracy is ensured based on the deep network, and convenience is provided for the user to watch videos.

The invention discloses a video similarity detection method, and the method comprises the following steps: determining a to-be-detected video frame in a to-be-detected video when the to-be-detected video is obtained; screening out a to-be-detected video segment from the to-be-detected video by using the to-be-detected video frame; matching a selected preset video segment meeting a preset condition with the to-be-detected video segment in a preset video library with the to-be-detected video segment to obtain a video pair; and performing similarity detection on the video pair to obtain a similarity detection result of the to-be-detected video. The invention further discloses a video similarity detection device, terminal equipment and a computer readable storage medium. By using the video similarity detection method provided by the invention, the video similarity detection efficiency is improved.

The invention discloses a method for supplementing a public security investigation data set based on a selfie buffeting view. The method includes: utilizing a customized interface to dock a short video content service platform to obtain network short video; receiving, storing and establishing a network short video library according to view library criteria,; extracting target information by utilizing an AI analysis system according to a view library standard; forming a short video view database; and restoring beautified target images such as faces and human bodies by using an adversarial beautification technology, supplementing the images to a short video view database, extracting target information from the beautified and restored target images such as faces and human bodies by using secondary AI analysis, supplementing the target information to the short video view database, and performing fusion multi-dimensional analysis by using public security big data. According to the method, data mining of the short video view data is realized.

Embodiments of the disclosure provides methods and apparatuses for video processing. In one embodiment, the video processing method comprises: obtaining at least one video from a video repository as a video to be processed; performing semantic recognition on the video in one or more semantic recognition dimensions to obtain one or more video label data items corresponding to the video in the one or more semantic recognition dimensions; generating at least one candidate label combination based on at least one of the one or more video label data items; determining, based on a target label combination selected by a user from the at least one candidate label combination, one or more video clips in the video corresponding to at least one video label in the target label combination; and generating at least one target video clip corresponding to the target label combination based on at least one of the one or more video clips.

The embodiment of the invention discloses a similar video screening method and device, equipment and a storage medium. The method comprises the following steps: constructing a similar video library ofeach to-be-selected video under a specified similar scale according to video-level characteristics of each to-be-selected video in a to-be-selected video library; searching similar videos of which the similarity with the to-be-selected video exceeds a preset similarity threshold from the similar video library of each to-be-selected video to obtain a corresponding video pair candidate library; andscreening out corresponding similar video pairs from the video pair candidate library based on the frame-level features of the two video parties of each candidate video pair in the video pair candidate library under the corresponding similarity weights. According to the technical scheme provided by the embodiment of the invention, the comprehensiveness and accuracy of similar video screening areimproved by adopting a double-screening mode, and meanwhile, the extraction operand of the frame-level features of the two video parties is controlled, so that the extraction operand of the frame-level features is reduced, and the screening efficiency of similar videos is improved.

The invention provides a video association method and device, a server and a readable storage medium, and belongs to the technical field of computers. The method comprises the steps of obtaining firstdescription text information corresponding to a target short video; acquiring a plurality of long videos matched with the target short video in a preset long video library, wherein the long video inthe preset long video library has corresponding second description text information; obtaining a plurality of long videos matched with the target short video based on the first description text information and the second description text information; and obtaining a target long video associated with the target short video from the plurality of long videos based on the video similarity between thetarget short video and the plurality of long videos. Through the scheme, the plurality of long videos associated with the target short video can be matched by describing the text information, and thenthe target long video with high association degree with the target short video is matched in the long videos through the AI association analysis algorithm, so that the AI processing amount can be effectively reduced, and the AI resources are saved.

The invention discloses a video comment generation method and device based on artificial intelligence, an electronic device and a computer readable storage medium. The method comprises the steps of determining a similar video set similar to a to-be-commented video in a video library; obtaining a historical comment set of the similar video set; and generating video comments of the to-be-commented video based on the historical comment set of the similar video set. According to the method, a video publisher can be stimulated to continue to create and publish videos, so that the creation atmosphere of a platform is more benign, and the community operation atmosphere of the video platform is activated.