844results about "Video data clustering/classification" patented technology

The invention discloses a video classification method and device and a server. The method comprises the following steps of: obtaining a target video; The image frames in the target video are classified by the first classification model, and the image classification result is obtained. The first classification model is used for classification based on the image features of the image frames. The audio in the target video is classified by the second classification model, and the audio classification result is obtained. The second classification model is used to classify the audio based on the audio features. The text description information corresponding to the target video is classified by the third classification model, and the text classification result is obtained. The third classification model is used to classify the text information based on the text characteristics of the text description information. According to the image classification results, audio classification results andtext classification results, the target video target classification results are determined. In the present application, image features, audio features and text features are integrated for classification, and features of different dimensions of the video are fully considered, thereby improving the accuracy of the video classification.

The present invention relates to the field of security technology, and in particular to a fine granularity real-time supervision system based on edge computing. A fine granularity real-time supervision system based on edge computing is provided, comprising: an intelligent video monitoring device, an edge computing module, an edge computing node, and a cloud computing center. The system can reduce the redundant information of the system and realizes fine granularity management.

Techniques for printing stored or recorded information that may include information of one or more types and may be stored in one or more multimedia documents. A graphical user interface (GUI) is provided that facilitates printing of information stored by the one or more multimedia documents or portions of the information stored by the one or more multimedia documents.

A system, method and computer program product for extracting metadata from one or more content files in faster than real-time, where the content files may be received from more than one source.

A method for identifying object images in a display of a video signal, storing object identification information in an overlay directory, and extracting object information relevant to identified objects from respective information databases is disclosed. Overlay directories from several video signals and associated object information are assembled in an overlay server accessed by computers directly or through a telecommunication network. Object images are selected by users of authoring computers accessing the overlay server. An overlay directory for a video recording indicates temporal and spatial coordinates of images of selected objects, a network address of a video source providing the video recording, descriptions of the selected objects, and addresses of servers providing detailed information of the selected objects. The objects may include commercial products, services, or educational material. The disclosed method therefore may be used for advertising or general information dissemination purposes.

The disclosure provides methods and systems for interactive video content delivery. An example method comprises receiving a video content such as live television or video streaming. The method can run one or more machine-learning classifiers on video frames of the video content to create classification metadata corresponding to the machine-learning classifiers and one or more probability scores associated with the classification metadata. Furthermore, the method can create one or more interaction triggers based on a set of predetermined rules and optionally user profiles. The method can determine that a condition for triggering at least one of the triggers is met and triggers at least one of the actions with regard to the video content based on the determination, the classification metadata, and the probability scores. For example, the action can deliver additional information, present recommendations, automatically edit the video content, or control delivery of video content.

An exemplary virtual reality media system presents a field of view of an immersive virtual reality world on a display screen of a media player device associated with a user. The field of view includes content of the immersive virtual reality world and dynamically changes in response to user input provided by the user as the user experiences the immersive virtual reality world. Additionally, the virtual reality media system detects that a gaze of the user is directed for a predetermined amount of time at a gaze target included within the field of view. In response to the detection, the virtual reality media system presents an interactive user interface associated with the gaze target. In some examples, the interactive user interface is presented within the field of view together with the content of the immersive virtual reality world. Corresponding methods and systems are also described.

The invention relates to an intelligent education video service system based on cloud computing and a mobile terminal and an operation method thereof. The intelligent education video service system comprises a multi-platform terminal, a communication network, a front-end processing module, a short video module, a first intelligent auditing module, a live broadcast module, a second intelligent auditing module, an interactive learning and testing module, a data analysis module and a cloud computing module. The short video module has strong capabilities of rapid uploading, transcoding, storage, distribution and the like, integrates functions of shooting, special effects, filtering, editing, synthesis, local compression, uploading, playing, real-time interactive test questions and the like, and realizes editing, manufacturing and publishing of one-stop education short videos. The live broadcast module displays auxiliary information such as student answering statistical percentage and answering time in real time, and teachers dynamically adjust teaching plans or explain and consolidate detailed knowledge points according to displayed real-time statistical results.

The invention relates to a video classification processing method and device, computer equipment and a storage medium. The method comprises the steps of obtaining a target video; extracting multi-modal data of the target video; performing classification prediction on the data of each mode to obtain probability vectors which respectively correspond to each mode and are used for representing the probability that the target video belongs to each preset category; combining the probability vectors corresponding to the modalities; and predicting the final category of the target video according to the combined vectors. According to the scheme, the accuracy of video classification can be improved.

According to one embodiment of the present invention, a computer implemented method, apparatus, and computer-usable program product for generating unique cohort groups using multimodal sensory device. Multimodal sensory data is received from a set of multimodal sensors in a public environment. The set of multimodal sensors are associated with a network. The multimodal sensory data is received from the set of multimodal sensors over the network. The multimodal sensory data is processed to generate a plurality of attributes to form cohort attributes. A plurality of unique cohort groups is generated using the cohort attributes and the multimodal sensory data. Each member of the cohort group shares at least one common attribute.

The embodiment of the invention discloses a model training method, a video category detection method and device, an electronic device and a computer readable medium. An embodiment of the video category detection method comprises the steps of extracting a key frame of a target video, and generating a key frame sequence; inputting the key frame sequence into a feature extraction model to obtain a feature information sequence corresponding to the key frame sequence; and inputting the feature information sequence into a video category detection model to obtain a category detection result of the target video. According to the method and the device, the video category detection efficiency is improved.

A method and apparatus for retrieving information relevant to tracked objects appearing in a display of a video signal is disclosed. The method is performed by a viewing computer having stored thereon an augmented display tool. In response to a user requesting the video signal, a content directory storing content information relevant to the tracked objects is acquired from a video-overlay server. The augmented display tool causes the viewing computer to acquire and display the video signal and record a time measurement and spatial coordinates of each point selected by a viewer using a pointing device. The augmented display tool uses the content directory to find an object identifier corresponding to each selected point and extracts relevant information from a global object directory maintained at the video-overlay server.

The embodiment of the invention discloses a video classification method. The method comprises the steps of obtaining a key frame image from a target video; inputting the key frame image into an imagesearch engine to obtain description information of the key frame image, and determining a keyword group of the key frame image according to the description information; obtaining text content characteristics corresponding to the keyword group; and determining a video type label of the target video according to the text content characteristics. By adopting the method and the device, the text content characteristics of the target video can be determined on the basis of the key frame images in the plurality of frame images forming the target video and the corresponding description information inthe image search engine, so that the video type label of the target video is obtained, and the video classification efficiency is improved.

The invention relates to a content classification method and a device, a computer device and a storage medium. The method comprises the steps of obtaining a target feature vector corresponding to to-be-classified target content; obtaining a target classification model obtained by training, wherein the target classification model comprises a first classification model and a second classification model; inputting the target feature vector into a first classification model to obtain a first content category corresponding to the target content, the first content category being a content category corresponding to the first classification hierarchy; obtaining first category feature information corresponding to the first classification hierarchy; inputting the first category feature information and the target feature vector into a second classification model to obtain a second content category corresponding to the target content, the second content category being a content category corresponding to a second classification level, and the level of the second classification level being lower than the level of the first classification level; and taking the first content category and the second content category as classification results corresponding to the target content. The method can improve the content classification accuracy.

The invention relates to a double-stream video classification method and device based on a cross-modal attention mechanism. The method is different from a traditional double-flow method. Information of two modes (even more modes) is fused before a prediction result. Therefore, the method is more efficient and sufficient. Meanwhile, information interaction is carried out in the earlier stage, a single branch has important information of another branch in the later stage, the precision of the single branch is flush with or even exceeds that of a traditional double-flow method, and the parameterquantity of the single branch is much less than that of the traditional double-flow method. Compared with a non-local neural network, the attention module designed by the invention can be in a cross-mode state instead of only using an attention mechanism in a single mode, and the effect of the method provided by the invention is equivalent to that of the non-local neural network under the condition that the two modes are the same.

The invention provides a video clip tag identification method. The method comprises the steps of obtaining a target video clip; extracting image features and audio features of the target video clip; and analyzing image features and audio features of the target video clip by using a pre-trained multi-label classification model to obtain a tag classification result of the target video clip, the label classification result of the target video clip comprising category labels of the target video clip in at least two dimensions. Based on the scheme provided by the invention, the tag of the target video clip can be comprehensively identified, and the accuracy of a tag identification result can be improved.

The present invention provides an easy-to-use centralized service for providing and using advanced video and audio browsing and tagging methods to create a revised and improved video media set and for enabling a user to auto-create a fixed media form of the so-edited and so-improved video. The present invention also enables a system that allows users to select varying degrees of automated creation of a fixed media form recording following editing and revision steps potentially involving synchronized tagging and commenting aspects. Systems and operational modes are provided for labeling and formatting the auto-generated fixed media data.

The invention relates to a video classification method, device and equipment, and a computer readable storage medium. The method comprises the steps of obtaining at least two key frame picture segments of a to-be-classified video; according to the time-space information of the at least two key frame picture segments, segment semantic vectors of the at least two key frame picture segments are acquired respectively; performing bidirectional association fusion on the at least two fragment semantic vectors to obtain a global to-be-classified semantic vector of the to-be-classified video; obtaininga prediction probability of the to-be-classified video according to the global to-be-classified semantic vector; wherein the prediction probability is used for determining a classification result ofthe to-be-classified video. By adopting the method, the video classification accuracy can be effectively improved.

The invention discloses a method and device for generating user characteristics, equipment and a storage medium and relates to the field of video recommendation. The method for generating the user characteristics comprises the steps of: obtaining a time series corresponding relationship between n groups of target videos and user accounts to obtain a word embedding matrix; training the word embedding matrix by using a loss function; and determining word vectors corresponding to the user accounts in the trained word embedding matrix as the user characteristics of the user accounts. According tothe method, the user characteristics are extracted on the basis of watching history and / or searching history of users; as long as the users normally use a video system, the data are continuously generated and updated without relying on other data sources; and therefore, the problem that the effective user characteristics cannot be generated for the users with empty or incomplete or incorrect attribute information, of a method in related work, can be solved and the accurate user characteristics can be generated for the users who use the video system.

The invention discloses a method for generating a video by extracting a multimedia material based on a template. The method comprises the following steps: acquiring multimedia materials, preprocessingthe multimedia materials, performing tagging processing, outputting tags of the multimedia materials, and clustering the multimedia materials and the corresponding tags according to a preset clustering rule to obtain a plurality of data sets; obtaining template configuration data input by a user, and establishing a video template according to the template configuration data and a preset initial template; and automatically performing a video generation task according to the template configuration data, extracting the multimedia material according to the template configuration data and the label, generating a video according to the template configuration data and the extracted multimedia material, and outputting the video. According to the invention, the user do not need to search, screen and confirm video materials, and video generation tasks can be automatically carried out according to user requirements, thereby reducing repeated operation processes of the users.

The embodiment of the invention provides a video collection generation method and device, electronic equipment and a storage medium. The method comprises the following steps: classifying according tothe content of a video frame to obtain a classification result; Determining a specific video segment containing predetermined content according to the classification result; And generating a video collection based on the specific video segment.

A computer-implemented method for managing video contents includes collecting a plurality of keywords related to a topic, the keywords being collected using at least one dynamic data source. One or more sub-topics of the topic are identified using the keywords collected. A topic node in a taxonomy of topics is built, the topic node including a topic identifier for the topic, a child topic identifier for the sub-topics identified, and a keyword section for one or more of the keywords collected. A plurality of videos is organized using the topic node built to assist a user in locating a video of interest.

A display apparatus including a display, a communication unit configured to communicate with a plurality of terminal devices and receive image contents from the plurality of terminal devices, a storage configured to store the received image contents, and a processor configured to classify the stored image content according to a predetermined criterion and display the image contents through the display.

A computer-implemented method for managing video contents includes collecting a plurality of keywords related to a topic, the keywords being collected using at least one dynamic data source. One or more sub-topics of the topic are identified using the keywords collected. A topic node in a taxonomy of topics is built, the topic node including a topic identifier for the topic, a child topic identifier for the sub-topics identified, and a keyword section for one or more of the keywords collected. A plurality of videos is organized using the topic node built to assist a user in locating a video of interest.

In order to solve the problem of small relevance between target video clip extraction and user interestingness, the invention provides a target video clip extraction method, which comprises the following steps of: obtaining target scene classification information and target person information; identifying a scene to obtain a scene classification wonderful degree score of each video clip; identifying the character to obtain a target character identification wonderful degree score of each video clip; generating a user option wonderful degree score according to the scene classification wonderfuldegree score and the target character identification wonderful degree score; according to the image difference, obtaining an image wonderful degree score of the video clip; according to the short-timeenergy value of the audio frame, acquiring an audio highlight score of each video clip; obtaining a content highlight score according to the image highlight score and the audio highlight score; obtaining a diversity score between the video clips according to the distance between the video clips; and according to an optimization objective function f(X)=[Sigma]i<XUScore(i)*w1+[Sigma]i<XCScore(i)*w2+[Sigma]i, j<XDScore(i, j)*w3, selecting a preset value Nsel video clips to form a target video clip set X, so that the value of f(X) is maximum.

Systems and methods for compressing and sharing geotagged video in accordance with embodiments of the invention are disclosed. One embodiment includes receiving a captured video sequence, where at least one geographic location is associated with the captured video sequence, selecting a segment of the captured video sequence, identifying a set of relevant video segments from a geotagged video database based on the at least one geotag associated with the captured video sequence, determining the video segment from the set of relevant video segments that is the best match by comparing the similarity of the content in the video segments to the content of the selected segment from the captured video sequence, encoding the selected segment, where the selected segment is encoded using predictions that include references to the video segment that is the best match, and storing the encoded video segment in the geotagged video database.

The invention provides a video classification method and device, electronic equipment and a computer readable storage medium, and relates to the technical field of image processing. The video classification method comprises the steps: performing sparse sampling on a to-be-processed video to obtain a plurality of key frames; processing the plurality of key frames through a feature extraction network in a preset model to extract features of the plurality of key frames; and fusing the features of the plurality of key frames through a trained attention network in the preset model, and processing the fused features to obtain a classification result of the to-be-processed video. The video classification method can reduce the calculation amount, and improves the video classification speed and efficiency.

The invention discloses a video auditing method and device, an auditing server and a storage medium. The method comprises the following steps: inputting each key video frame in a to-be-audited video into a violation classification model, and obtaining a classification score of each key video frame under a corresponding violation sub-class through a binary classification module under different violation sub-classes in the violation classification model; for each violation sub-category, fusing the classification score of each key video frame under the violation sub-category to obtain a violationscore of the to-be-audited video under the violation sub-category; and determining violation category composition of the to-be-audited video according to the violation scores of the to-be-audited video under different violation sub-categories and a preset violation threshold. According to the technical scheme provided by the invention, multi-violation category judgment of the to-be-audited videois realized, the problem of misjudgment or missed judgment of violation sub-categories is avoided, the audit independence of the to-be-audited video under different violation sub-categories is ensured, and the comprehensiveness and accuracy of video audit are improved.