20 results about "Video learning" patented technology

The invention discloses an artificial intelligence robot system based on human video learning, which comprises a human operation video data set and a small-scale robot teaching data set, and is characterized in that human operation videos of various daily tasks are stored in the human operation video data set; a small amount of robot teaching data is stored in the small-scale robot teaching data set, the output end of the human operation video data set is in signal connection with a general operation prediction model, and the general operation prediction model is used for learning physical interaction rules according to human operation videos in the human operation video data set. And the output end of the general operation prediction model is in signal connection with a physical interaction rule generation module. According to the method, the error-prone steps of attitude estimation, redirection and image restoration are completely abandoned, knowledge migration is directly carried out on advanced semantics and representation levels, the learning precision is ensured, and the conversion link of error accumulation is avoided.

The application discloses a method for optimizing a sign language library, comprising the following steps: providing a sign language library with built-in basic data tables; providing a user data table in the sign language library, which is used for storing emotional sign language data with emotional types; obtaining a sign language sample provided by a target user, the sign language sample comprising a sample word, a sample emotional type and a sample sign language video; learning sample sign language actions and sample action execution time according to the sample sign language video; updating the user data table according to the sample word, the sample emotional type, the sample sign language actions and the sample action execution time; wherein, when providing a sign language library service, the user data table has a high priority, and the basic data table has a low priority. The technical scheme provided by the application can enable a user to match correct semantics in time and convey emotions at a lower cost.

This invention relates to the fields of wireless sensing and deep learning, specifically to a human pose estimation method based on a multimodal attention network. In video modality learning, a multi-resolution network (FCN) is designed to extract pose features. By applying a local self-attention network to the generated heatmap, the weights and attention levels of different regions are adaptively adjusted to capture more refined keypoint information. In CSI modality learning, a spatiotemporal attention network and a multimodal-guided linear spatial position variation layer are designed to facilitate self-learning of its spatiotemporal features. A teacher-student architecture is adopted in the overall multimodal learning network to allow the Wi-Fi signal-based pose estimation model to learn more from the deep learning capabilities of the video modality. By transferring the correct human pose estimation information from the video learning network to the Wi-Fi signal learning network, accurate human pose information can be estimated using only the CSI modality as input.

The application relates to a sparse video adversarial attack method and device based on a space-time reversible neural network. The method comprises the following steps: acquiring a guide video of a target class; learning a learnable target feature tensor under the guidance of the guide video of the target class and a guide target classification model by adopting a guide target video learning module; and performing space-time feature space information exchange between the target feature tensor learned by the guide target video learning module and a clean video by adopting a space-time reversible neural network module to generate an adversarial video and a residual video; the space-time reversible neural network module is used for utilizing the information preservation characteristic of the reversible neural network, performing information exchange between the target feature tensor and the clean video at a space-time feature level under the drive of an adversarial loss function, and generating the adversarial video and the residual video. The method fully considers the space-time features of the video, and generates a robust adversarial video sample which is difficult to detect.

1. Name of the product in this design: Video Learning Graphical User Interface (Focus) for Electronic Devices. 2. Purpose of this design: to run programs and display information. 3. The key design features of this product are its graphical user interface content. 4. The image or photograph that best illustrates the design's key features: the front view. 5. Electronic devices using this graphical user interface are of conventional design, so the rear view is omitted; electronic devices using this graphical user interface are of conventional design, so the left view is omitted; electronic devices using this graphical user interface are of conventional design, so the right view is omitted; electronic devices using this graphical user interface are of conventional design, so the top view is omitted; electronic devices using this graphical user interface are of conventional design, so the bottom view is omitted. 6. Purpose of the graphical user interface: to display relevant information about the video learning app to users and to enable human-computer interaction. 7. Other situations requiring explanation are illustrated in the diagrams: The product's graphical user interface is the operating interface of the learning robot, and the main view display interface is the main interface of the video learning APP after the user has successfully logged in; Clicking the "Peak Scholar" icon in the middle tab bar of the main view interface leads to interface change state diagram 1; Clicking the "Focus" icon in interface change state diagram 1 leads to interface change state diagram 2; Clicking the "Start Training" icon in interface change state diagram 2 leads to interface change state diagram 3; Clicking the "Attention Training" icon in interface change state diagram 3 leads to interface change state diagram 4; Clicking the "Start" icon in interface change state diagram 4 leads to interface change state diagram 5; After waiting 60 seconds, it automatically enters interface change state diagram 6; After waiting another 60 seconds, it automatically enters interface change state diagram 7; After waiting another 60 seconds, it automatically enters interface change state diagram 8; After waiting another 60 seconds, it automatically enters interface change state diagram 9; The upper left gray area of ​​the main view, interface change state diagram 1, interface change state diagram 2, and interface change state diagram 3 represents variable content.

The invention discloses a video learner attention intelligent prediction method based on a decoupling type space-time state space model. The method comprises the following steps: acquiring a learning content video and uniformly sampling multiple frames; initial features of each frame are extracted, and spatial enhancement features are obtained through a spatial module; constructing a position-level time sequence according to spatial position alignment, and obtaining a time enhancement feature through a time module; splicing the space and time enhancement features, and performing linear projection fusion to obtain space-time enhancement features; decoding and outputting a frame-by-frame attention saliency map, and intelligently generating a learner video attention area; and the training adopts a composite target optimization model parameter. According to the invention, the attention area of the video learner can be perceived in advance and intelligently predicted, the key area of the teaching video is optimized and adjusted according to the prediction result, and the learning effect is improved. In a word, compared with the prior art, spatial feature modeling and time feature modeling are decoupled, and the method has the advantages of being good in prediction effect, high in intelligent level and outstanding in education scene application value.

The system according to this embodiment aims to analyze the activity status and physical condition of pets in real time and provide them with an optimal environment. [Solution] The system according to the embodiment comprises an analysis unit, an environment optimization unit, a learning unit, and a nurturing unit. The analysis unit analyzes the pet's activity status and physical condition in real time. The environment optimization unit provides the optimal environment based on the data analyzed by the analysis unit. The learning unit learns the pet's behavior based on the video footage from a camera set up in the room. The nurturing unit allows the pet to live with you indefinitely on your smartphone.

1. The name of the design product: video learning graphical user interface (camera memory) for electronic device. 2. The use of the design product: running programs, displaying information. 3. The design points of the design product: the interface content of the graphical user interface. 4. The picture or photo that best indicates the design points: front view. 5. The electronic device applying the graphical user interface is of conventional design, omitting the rear view; the electronic device applying the graphical user interface is of conventional design, omitting the left view; the electronic device applying the graphical user interface is of conventional design, omitting the right view; the electronic device applying the graphical user interface is of conventional design, omitting the top view; the electronic device applying the graphical user interface is of conventional design, omitting the bottom view. 6. The use of the graphical user interface: for showing video learning APP related information to the user and realizing human-computer interaction. 7. Other circumstances that need to be explained: the graphical user interface of the product is the operation interface of a learning robot, the front view display interface is the main interface after the user of the video learning APP logs in successfully; clicking the "peak scholar" icon in the middle of the tab bar in the front view interface enters interface change state diagram 1; clicking the "camera memory" icon in interface change state diagram 1 enters interface change state diagram 2; clicking the "training plan 1" icon in interface change state diagram 2 enters interface change state diagram 3; clicking the "overall perception / enlarged field of view" icon in interface change state diagram 3 enters interface change state diagram 4; clicking the "start" icon in interface change state diagram 4 enters interface change state diagram 5; waiting for 60 seconds automatically enters interface change state diagram 6; waiting for another 60 seconds automatically enters interface change state diagram 7; waiting for another 60 seconds automatically enters interface change state diagram 8; waiting for another 60 seconds automatically enters interface change state diagram 9; waiting for another 60 seconds automatically enters interface change state diagram 10; waiting for another 60 seconds automatically enters interface change state diagram 11; waiting for another 60 seconds automatically enters interface change state diagram 12; waiting for another 60 seconds automatically enters interface change state diagram 13; waiting for another 60 seconds automatically enters interface change state diagram 14; waiting for another 60 seconds automatically enters interface change state diagram 15; waiting for another 60 seconds automatically enters interface change state diagram 16; waiting for another 60 seconds automatically enters interface change state diagram 17; waiting for another 60 seconds automatically returns to interface change state diagram 4; the left upper gray block area of the front view, interface change state diagram 1, interface change state diagram 2 and interface change state diagram 3 is a variable content screen.

The present specification is developed about high-robustness bodybuilding robot systems and methods. The system is rich in module, and the basic module covers the functions of perception, action planning, speech synthesis, emotion coordination, role management and the like; the additional module has the characteristics of privacy protection, emotion interaction, role switching, physiological simulation, sound cloning, remote interaction, transfer prevention, cleaning and lubrication, video learning, organ replacement and the like. According to the demonstration and personalized extension method, demonstration action tracks and multi-modal data are collected, sub-fragments are decomposed to generate new track elements, the new track elements are stored, and meanwhile a semantic-action association database is established. Based on the method and the system, permanent expansion of personalized intimate interaction actions can be realized, new skills can be continuously learned according to different user requirements, accompanying services highly fitting expectations are provided for users, and practicability and user satisfaction are improved and multiple personalized requirements are met in multiple scenes such as emotion accompanying and life assistance.

1. Name of the product in this design: Video Learning Machine (R3). 2. Purpose of this design: This product is a tablet audio and video player used for playing audio files. 3. The key design features of this product are: 1) The key design feature of this product is that the overall shape is flat. 2) The key design feature of this product is the movable support on the back. 3) The key design feature of this product is that there are speaker holes on both sides. 4. The image or photograph that best illustrates the design's key features: the front view.

This invention relates to the field of video pattern recognition technology, specifically to an intelligent video learning engine and its multi-level real-time knowledge point localization method. The invention first obtains the change coefficient of each pixel in each frame based on the changes in image content between adjacent frames of the learning video; further, it obtains the real-time change distance between pixels based on the temporal fluctuation differences of the change coefficients and performs iterative windowing; based on the distribution of the change coefficients of pixels in the real-time image, combined with a synchronization factor, the learning video is segmented; further, based on the continuity of the text content, video segmentation points are removed; finally, an interactive knowledge point tree is constructed based on the video segmentation results and the text content. This invention, by integrating image change analysis and text semantic continuity verification, intelligently divides teaching video segments and ultimately constructs an interactive knowledge point tree, achieving accurate localization and efficient navigation of learning content.

The invention provides an intelligent management system for semiconductor equipment, and the system comprises a sensing layer which is used for obtaining the sensing information of the equipment, and the sensing information comprises the operation state, environment parameters, alarm, maintenance history, and equipment fault pictures or videos of the equipment; the learning and reasoning layer is used for processing, calculating and analyzing perception information through learning and reasoning capabilities of a large model and a knowledge base; the decision-making layer is used for equipment data analysis decision-making, equipment fault intelligent diagnosis and scheme pushing; and the execution layer is used for executing the decision-making result made by the decision-making layer and performing corresponding execution actions. The semiconductor equipment intelligent management system integrates, analyzes and excavates information such as equipment state, alarm, maintenance history, equipment fault pictures / videos and the like, combines big data, artificial intelligence and knowledge base technologies, constructs an equipment management and maintenance agent, predicts equipment faults, reduces the equipment fault rate, improves the equipment comprehensive efficiency, and reduces the maintenance cost. The working efficiency of equipment engineers is improved, and comprehensive intelligentization of equipment management is realized.

The invention belongs to the crossing field of artificial intelligence and medical information technology, and relates to a method and system for improving the expandability of a large language model in low-AI health vegetarian population. The method comprises the following steps: constructing a main chat robot, and obtaining dialogue log data of a user and the main chat robot; extracting an interaction obstacle type list according to the dialogue log data; constructing an interaction obstacle-learning instruction mapping form according to the interaction obstacle type list; generating video learning content for each learning instruction according to the interaction obstacle-learning instruction mapping form; and adaptively distributing the video learning content, and accurately putting the video learning content to a specified interface position of the target user. According to the method, the dialogue ability of a large model and a patient can be improved, and the expandability problem of low-AI healthy vegetarian people is solved.

An object of the system according to the embodiment is to automatically extract a specific scene from a game video.SOLUTION: A system according to an embodiment includes a tutorial video learning unit, a game video analysis unit, and a scene extraction unit. The tutorial video learning unit learns a tutorial video. The match video analysis unit analyzes the match video on the basis of the feature learned by the tutorial video learning unit. The scene extraction unit extracts a scene matching the learned feature from the game video analyzed by the game video analysis unit.SELECTED DRAWING: Figure 1

1. The name of the design product: video learning graphical user interface (quick reading) for electronic device. 2. The use of the design product: running programs, displaying information. 3. The design points of the design product: the interface content of the graphical user interface. 4. The picture or photo that best indicates the design points: front view. 5. The electronic device applying the graphical user interface is of conventional design, and the rear view is omitted; the electronic device applying the graphical user interface is of conventional design, and the left view is omitted; the electronic device applying the graphical user interface is of conventional design, and the right view is omitted; the electronic device applying the graphical user interface is of conventional design, and the top view is omitted; the electronic device applying the graphical user interface is of conventional design, and the bottom view is omitted. 6. The use of the graphical user interface: for showing the user the relevant information of the video learning APP in the electronic device and realizing human-computer interaction. 7. Other circumstances that need to be explained: the graphical user interface of the product is the operation interface of a learning robot, and the front view shows the main interface after the user of the video learning APP logs in successfully; clicking the "peak scholar" icon in the middle of the label bar in the front view enters interface change state diagram 1; clicking the "quick reading" icon in interface change state diagram 1 enters interface change state diagram 2; clicking the "start training" icon in interface change state diagram 2 enters interface change state diagram 3; clicking the "reading rhythm" icon in interface change state diagram 3 enters interface change state diagram 4; clicking the "start" icon in interface change state diagram 4 enters interface change state diagram 5; waiting for 60 seconds automatically enters interface change state diagram 6; waiting for another 60 seconds automatically enters interface change state diagram 7; waiting for another 60 seconds automatically enters interface change state diagram 8; waiting for another 60 seconds automatically enters interface change state diagram 9; waiting for another 60 seconds automatically enters interface change state diagram 10; waiting for another 60 seconds automatically returns to interface change state diagram 4; the left upper gray block area of the front view, interface change state diagram 1, interface change state diagram 2, and interface change state diagram 3 is a variable content screen.

The invention relates to the technical field of online education recommendation, in particular to a personalized video learning recommendation method fusing hierarchical reinforcement learning and a knowledge graph, and the method comprises the steps: obtaining video data and a user portrait vector; wherein the video data comprises video resources and knowledge points associated with the video resources; constructing a target knowledge graph based on the video resources and the knowledge points associated with the video resources, and generating a knowledge graph embedding vector based on a graph neural network; and generating a personalized recommendation result of the user through a knowledge grouping aggregation mechanism and an attention mechanism based on the user portrait vector and the knowledge graph embedding vector, and generating a self-adaptive learning path of the user based on the user portrait vector and a target knowledge graph when the user selects a learning target. According to the method, the technical problems of noise historical interference, weak knowledge relevance, inaccurate path planning and the like in a traditional recommendation system can be solved.

An object of a system according to an embodiment is to automatically execute work of a user.SOLUTION: A system includes a moving image analysis unit, a learning unit, and an automatic operation unit. The video analysis unit analyzes the video of the screenshot. The learning unit learns the work of the user from the moving image of the screen shot analyzed by the moving image analysis unit. The automatic operation unit automatically executes a mouse operation or a keyboard operation based on the work learned by the learning unit.SELECTED DRAWING: Figure 1

Little-vision learning of repetitive human tasks is provided. Little-vision learning of repetitive human tasks is performed. Sliding window-based temporal segmentation of sensor data is performed for multiple cycles of a repetitive task. Motion alignment of the multiple cycles is performed that maps portions of the multiple cycles to corresponding portions of other cycles of the multiple cycles. Classes are constructed for each corresponding portion of the multiple cycles according to the motion alignment. Meta-training is performed to teach a model from data sampled from a set of labeled human motions and the classes of each corresponding portion, the model utilizing a bidirectional long short-term memory (LSTM) network to account for length variations between the multiple cycles. The model is used to perform temporal segmentation of a data stream of sensor data in real-time for predicting motion windows within the data stream.

The application discloses a video learning emotion recognition method based on emotion infection tracking and multi-modal fusion, which comprises the following steps: 1) data collection; 2) data preprocessing; 3) feature structure enhancement; 4) emotion infection driven feature enhancement and emotion recognition; 5) testing and evaluation. This method uses eye movement physiological signals to model the bias modulation of individual differences and emotional feedback of learners, and uses it as a guide to track the teaching video inducing factors and infect the feature weighting, and then realizes the deep enhancement fusion of physiological response and video features through the one-way emotion infection tracking module and the two-way cross-modal attention mechanism, finally improves the precision and robustness of emotion recognition.

The invention provides a video fine adjustment method, a video control method and an on-site quick disposal method, so that on one hand, the well-known problem that a target frame and a position in a video are difficult to quickly and accurately position by adopting an intelligent terminal of a traditional multi-point touch control technology is solved; therefore, no matter how the evidence video is expected to be quickly processed and extracted in video clipping depending on video learning and precision and on-site law enforcement of law enforcement officers, the technical restriction problem exists, and the man-machine interaction technology, the video fine adjustment method and the video control method provided in the disclosure are just used for solving the restriction problem. In addition, the invention also provides an on-site rapid disposal method, and solves the difficult problem of rapid evidence obtaining and video of current on-site event disposal such as rapid traffic accident disposal.