Patents
Literature
188 results about "Gesture recognition" patented technology
Efficacy Topic
Property
Owner
Technical Advancement
Application Domain
Technology Topic
Technology Field Word
Patent Country/Region
Patent Type
Patent Status
Application Year
Inventor
Gesture recognition is a topic in computer science and language technology with the goal of interpreting human gestures via mathematical algorithms. Gestures can originate from any bodily motion or state but commonly originate from the face or hand. Current focuses in the field include emotion recognition from face and hand gesture recognition. Users can use simple gestures to control or interact with devices without physically touching them. Many approaches have been made using cameras and computer vision algorithms to interpret sign language. However, the identification and recognition of posture, gait, proxemics, and human behaviors is also the subject of gesture recognition techniques. Gesture recognition can be seen as a way for computers to begin to understand human body language, thus building a richer bridge between machines and humans than primitive text user interfaces or even GUIs (graphical user interfaces), which still limit the majority of input to keyboard and mouse and interact naturally without any mechanical devices. Using the concept of gesture recognition, it is possible to point a finger at this point will move accordingly. This could make conventional input on devices such and even redundant.
A pose recognition method based on adaptive uncertainty-aware meta-learning
The application relates to the technical field of intelligent identification, and discloses a gesture recognition method based on adaptive uncertainty-aware meta learning, which formalizes a gesture small sample regression framework, extracts features by using a linearized neural network and a neural tangent kernel, establishes a Gaussian process regression model by combining Bayesian inference, projects a covariance matrix in a reduced dimension by using a Fisher information matrix in view of high-dimensional characteristics, simultaneously constructs an adaptive weight generator, generates a dynamic threshold value by means of historical loss moving average, training progress and an uncertainty correction term, and assigns specific weights to tasks; a posterior mean value is output as a predicted angle in a meta test stage, and the posterior covariance is used to quantize uncertainty. The application can effectively cope with object symmetry ambiguity and feature loss, focuses the model on difficult tasks through an adaptive mechanism, significantly improves prediction accuracy and reliability in a small sample scene, and reduces computational complexity.
Owner:NANKAI UNIV
Frame rate boost based on static gesture recognition
A method (1100) includes capturing, by a sensor (1002) in a computing device (1000), a first set of images (110A) at a first frame rate (1102); identifying a static gesture based on at least one image included in the first set of images (110A) (1104); in response to identifying the static gesture, capturing, by the sensor, a second set of images (110B) at a second frame rate, the second frame rate being faster than the first frame rate (1106); and identifying a dynamic gesture associated with the static gesture based on a plurality of images included in the second set of images (110B), the second set of images being captured after the first set of images (1108).
Owner:GOOGLE LLC
A smart ring with magnetic attraction charging and multi-functional interaction expansion and a matching terminal
PendingCN122250731ABatteries circuit arrangementsUser identity/authority verificationElectrical batterySimulation
This invention discloses a smart ring and its supporting terminal with magnetic charging and multi-functional interactive extension, belonging to the field of smart wearable devices and mobile terminal interaction technology. The back shell of the supporting terminal and the smart ring is provided with a concentric charging area, which is composed of a central strong magnet and two ring electrodes. The ring is D-shaped, with corresponding magnetic contacts on the flat section, and can be charged by adsorbing at any rotation angle. The convenient charging allows the ring to use a small battery, thereby integrating functions such as health sensing, gesture recognition, and security chip, and can also be used as a desktop stand and grip anchor to improve stability. Based on dual detection of physical contact and wireless communication, it also realizes extended interactive functions such as two-way anti-loss, dual identity authentication, and grip posture perception and mode switching. This invention effectively solves the problems of battery anxiety, easy loss, single function and space limitation of existing smart rings, and improves the integration of smart wearable devices and user experience.
Owner:GUANGZHOU INST OF RAILWAY TECH
A mobile phone screen unlocking method based on gesture recognition
The application relates to the technical field of mobile phone screen unlocking, and relates to a mobile phone screen unlocking method based on gesture recognition. The method comprises the following steps: S1, touch detection is performed on a mobile phone screen, so that a real-time touch pattern and a real-time completion time are generated; and S2, standard touch patterns and hand image data of set gesture types are collected. The application adjusts a time threshold value according to a formula by counting the average completion time of historical gesture operations of a user, so that the time threshold value is automatically lengthened if the actual operation time of the user is generally longer than the standard completion time, and the time threshold value is shortened if the actual operation time is shorter, the adjustment of the time threshold value avoids misjudgment caused by slow user operation speed, the unlocking is more in line with the actual operation rhythm of the user, and meanwhile, the basic deviation threshold value is adjusted according to the real-time completion time in the gesture comparison link, the adjustment avoids unlocking obstruction caused by temporary operation state fluctuation of the user, guarantees the rationality of judgment, and greatly improves operation convenience.
Owner:SHENZHEN SAIBO YUHUA ELECTRONIC TECH CO LTD
A cross-target gesture recognition method based on multi-modal fusion
The application relates to a multi-modal fusion cross-target gesture recognition method, which adopts a comparison fusion learning method to extract gesture features from WiFi data and video data, effectively solves the heterogeneity of the two modes, adopts a cross-modal generation method for the WiFi data, obtains WiFi data of a new user to solve the problem of missing target features caused by changes in the sensing target, and has good performance on a cross-target task and is superior to existing multi-modal fusion methods.
Owner:ZHONGBEI UNIV
Page transition display device and method and related product
PendingCN122086506AExecution for user interfacesInput/output processes for data processingComputer hardwareAnimation
The invention provides a page transition display device and method and a related product. The device comprises a gesture recognition module, a page management module, a UI frame module, a linked list management module, a model transformation module and a rendering display module. Wherein the page management module monitors gesture operation executed by a user in real time through the gesture recognition module, when a page transition event is recognized, a GPU instruction is generated for the page transition event through the UI frame module, and a first linked list and a second linked list are created and managed through the linked list management module on the basis of the GPU instruction. According to the method, image rendering can be carried out while 2.5 D conversion is executed in a mode of running double linked lists through double threads, so that the idle waiting time of a GPU (Graphics Processing Unit) is shortened, and the smoothness of page transition animation is improved.
Owner:ZHUHAI JIELI TECH
Animal mimicry processing method and device based on gesture recognition, equipment and medium
ActiveCN112766231Beasy to useEnhance interestBiometric pattern recognitionAnimal scienceComputer vision
The application discloses an animal mimic processing method and device based on gesture recognition, equipment and medium, the method comprises the following steps: acquiring image information of a mimic animal, and determining contour shape information of the image information of the mimic animal according to the image information of the mimic animal; according to the contour shape information, animal information corresponding to the contour shape information is matched out; according to the animal information corresponding to the contour shape information, an animal 3D model, sound and encyclopedic content corresponding to the animal information are automatically output and displayed.The application enables the intelligent terminal to have a new function, i.e., an animal mimic function based on gesture recognition; the image of the animal is mimicked through gestures, and the 3D animal image is presented to children through gesture recognition in the intelligent terminal, so that the interest and the inventiveness of children are improved through hands-on.
Owner:SHENZHEN SKYWORTH RGB ELECTRONICS CO LTD
A Method and System for Testing Gesture Recognition Performance Based on Extended Reality Terminals
PendingCN122131907AInput/output for user-computer interactionBiological modelsRobotic armComputer graphics (images)
This invention proposes a method and system for testing gesture recognition performance based on an extended reality terminal, relating to the field of human-computer interaction technology. The method includes: issuing gesture control commands to a bionic robotic arm; the bionic robotic arm simulating gestures using a reinforcement learning model and recording the gesture execution results according to the gesture control commands; detecting the gestures using an extended reality terminal and determining the gesture recognition results; capturing an image displayed on the extended reality terminal; and simultaneously analyzing the gesture execution results corresponding to the bionic robotic arm in the captured image and the gesture recognition results corresponding to the image displayed on the extended reality terminal to obtain gesture recognition performance test results.
Owner:CHINA ACADEMY OF INFORMATION & COMM
An event camera based low power gesture recognition method and system
The application belongs to the technical field of human-computer interaction, and particularly relates to a low-power-consumption gesture recognition method and system based on an event camera, which comprises the following steps: an event camera asynchronously monitors the brightness change of a pixel point in a field of view, and generates an event stream when the brightness change meets a preset threshold condition; meanwhile, a low-power-consumption microcontroller MCU keeps a working state, and a main application processor AP is in a sleep state; the application enables the low-power-consumption MCU to monitor the event stream all year round, and the main AP is only woken up when an effective gesture is recognized, so that the standby power consumption of the system can be as low as 10 mW. The light-weight filtering algorithm and the adaptive power consumption adjustment formula adopted by the MCU can dynamically optimize the power consumption according to the event rate; and the on-demand wake-up mechanism and the intelligent sleep strategy of the AP effectively avoid unnecessary energy waste.
Owner:SHANDONG ARTAPLAY INTELLIGENT TECH CO LTD
2d convolutional spatio-temporal excitation dynamic gesture recognition method based on contrastive learning enhancement
This invention provides a 2D convolutional spatiotemporally stimulated dynamic gesture recognition method based on contrastive learning enhancement, comprising: firstly, sampling and preprocessing a dynamic gesture video sequence to obtain multiple frames corresponding to the same gesture action; then, using a shared-parameter 2D convolutional neural network to extract spatial features frame by frame to obtain frame-level feature representations. Based on this, spatiotemporally stimulated enhancement of the frame-level features is performed through multi-scale temporal difference modeling, temporally adaptive weighted aggregation, and spatial attention stimulation to explicitly characterize the dynamic changes of the gesture action. The enhanced features within the same gesture sequence are constructed into a frame-level positive sample set, and a robust positive sample set center vector is calculated. Using this center vector as a positive anchor point, a contrastive learning mechanism based on hard negative sample weighting is introduced to constrain and optimize the feature space. During the training phase, the model parameters are optimized by jointly using classification loss and contrastive learning loss.
Owner:INST OF ADVANCED TECH UNIV OF SCI & TECH OF CHINA +1
Display device and dynamic gesture data labeling method
This application provides a method for labeling display devices and dynamic gesture data. The method includes: in the data acquisition stage, storing a gesture configuration file containing a specific execution order and time range to provide strong prior constraints for labeling and ensure that the acquired data meets expectations; secondly, using a gesture recognition model for preliminary labeling, and configuring a driving filtering algorithm to match the model inference results with the gesture configuration file to retain high-confidence labeled data; furthermore, introducing an uncertainty measure to distinguish between high- and low-value samples, directly adopting samples with low uncertainty, and manually correcting samples with high uncertainty to ensure the labeling accuracy of difficult samples; finally, verifying the labeling through multi-viewpoints by comparing the consistency of gestures in the same time period under different camera views to eliminate single-viewpoint errors and further improve the labeling accuracy, solving the problem of low labeling accuracy of current gesture video data and improving the accuracy of gesture video data labeling.
Owner:HISENSE ELECTRONICS TECH SHENZHEN CO LTD
Multi-level gesture recognition method and system for large screen interaction
PendingCN122261394Aimprove accuracyImprove effectivenessInput/output for user-computer interactionBiological modelsMedicineLarge screen
The application relates to the technical field of gesture interaction, in particular to a multi-level gesture recognition method and system for large-screen interaction, which solves the technical problem that in the prior art, due to mutual shielding of hands of multiple users in large-screen multi-user interaction, key points cannot be distinguished, and it is difficult to accurately complete and recognize user gestures. The method comprises the following steps: in a multi-user interaction scene, historical gesture data corresponding to each user and historical system operation instructions associated with the historical gesture data are acquired, and a gesture behavior model of each user is established; in an interaction process, when a target user is recognized, real-time gesture data of the target user is extracted, and the gesture behavior model of the target user is called; if missing of hand key points caused by shielding is found in the real-time gesture data of the target user, according to gesture data of the target user that is not shielded in a preset period before shielding occurs and the gesture behavior model of the target user, positions of the hand key points of the shielded part are predicted.
Owner:SHENZHEN MAZHE TECH CO LTD
Webxr gesture interaction method and device based on state machine
PendingCN122111216AInput/output for user-computer interactionGraph readingMedicineComputer graphics (images)
The application discloses a state machine-based WebXR gesture interaction method and device, which comprises the following steps: acquiring hand data of a user in real time through a WebXR device, performing gesture recognition on the hand data, and obtaining a gesture recognition result; inputting the gesture recognition result into a state machine, determining a target interaction state based on a current interaction state and the gesture recognition result through the state machine, and switching to the target interaction state; under the target interaction state, performing operation data calculation on the hand data according to a gesture mapping algorithm corresponding to the target interaction state, determining target 3D model operation data represented by the hand data; performing interactive operation on a target 3D model according to the model operation data, and driving a 3D rendering engine to update a visualized picture.
Owner:CHENGDU MEGAYOU TECH CO LTD
Context-sensitive control of radar-based gesture-recognition
ActiveUS12638925B2Input/output for user-computer interactionDigital data processing detailsRadar systemsProximity sensor
This document describes techniques and systems for radar-based gesture-recognition with context-sensitive gating and other context-sensitive controls. Sensor data from a proximity sensor and / or a movement sensor produces a context of a user equipment. The techniques and systems enable the user equipment to recognize contexts when a radar system can be unreliable and should not be used for gesture-recognition, enabling the user equipment to automatically disable or “gate” the output from the radar system according to context. The user equipment prevents the radar system from transitioning to a high-power state to perform gesture-recognition in contexts where radar data detected by the radar system is likely due to unintentional input. By so doing, the techniques conserve power, improve accuracy, or reduce latency relative to many common techniques and systems for radar-based gesture-recognition.
Owner:GOOGLE LLC
Palette control method and system for three-dimensional interactive scene
ActiveCN121095508BInput/output for user-computer interactionImage data processingComputer graphics (images)Engineering
The application is suitable for the technical field of three-dimensional drawing, and provides a palette control method and system for a three-dimensional interactive scene, comprising the following steps: receiving a three-dimensional model drawing instruction, entering a drawing surface selection mode, and receiving a drawing surface click instruction; identifying a model surface block where the click position is located, unfolding the model surface block into a plane to obtain a drawing plane; receiving a drawing instruction, displaying a drawing surface on the drawing plane, and mapping the drawing surface on a corresponding position on the model surface block; and performing control gesture recognition, automatically determining a control subject as a three-dimensional model or a drawing plane when a zoom gesture or a translation gesture is recognized, and automatically determining a rotation axis direction and a rotation axis position when a rotation gesture is recognized. The application converts complex three-dimensional surface drawing into relatively simple two-dimensional plane drawing, so that a user can more easily and accurately control the drawing of lines and patterns, and the accuracy and efficiency of drawing are improved.
Owner:HANGZHOU LONGYAO DIGITAL INTELLIGENCE TECHNOLOGY CO LTD
A target tracking method based on gesture recognition
This application discloses a target tracking method based on pose recognition. By acquiring the current image frame, pose recognition is performed on the detected target to obtain the coordinates and confidence scores of multiple pose keypoints. Valid keypoints are then selected based on the confidence scores, and a pose keypoint bounding box is constructed and expanded to obtain a pose-completed bounding box. Simultaneously, the orientation angle of the detected target is calculated. Based on this, a target trajectory is established and continuously updated based on previous image frames. For each target trajectory, a predicted trajectory position and predicted orientation angle corresponding to the current image frame are generated. The matching cost is calculated by combining the spatial relationship between the predicted trajectory position and the pose-completed bounding box. When multiple candidate detected targets exist, the final matched detected target is determined based on the orientation angle difference. This application improves the matching accuracy and tracking stability in occluded and crowded scenes through pose completion and orientation angle disambiguation mechanisms.
Owner:HEFEI JIANGXIN DUZHI INTELLIGENT TECH CO LTD
Split ar glasses for human-computer interaction
PendingCN122110514AInput/output for user-computer interactionGraph readingComputer graphics (images)Dextroscope
The application relates to the technical field of AR glasses, and discloses split AR glasses for human-computer interaction, which comprises glasses, the glasses comprising a glasses body, the two sides of the glasses body being connected with left and right glasses legs respectively, the glasses being electrically connected with a master controller, the middle top of the glasses body being provided with an RGB camera for collecting external scenes in real time, the two top sides of the glasses body being respectively provided with fisheye cameras for collecting user gesture images, display light machine modules being symmetrically arranged on the two sides of the glasses body of the RGB camera, and a nose pad being arranged on the back side of the glasses body and below the RGB camera. The glasses are highly integrated, small in device size, light in weight, powerful in function, and capable of realizing AR display, gesture recognition interaction, eye movement recognition interaction, real-time rear transmission of a field of view, space positioning and other functions. The glasses are comfortable to be attached to the head, good in experience, and good in man-machine ergonomics.
Owner:NAT INNOVATION INST OF DEFENSE TECH PLA ACAD OF MILITARY SCI
3D gesture recognition method, system, electronic device, storage medium and program product
PendingCN122135422AInput/output for user-computer interactionGraph readingComputer graphics (images)3d image
This invention provides a 3D gesture recognition method, system, electronic device, storage medium, and program product. The method includes: in a near-field interaction mode, acquiring a first 3D image containing a hand captured by a 3D camera; obtaining first 3D position information of the hand based on the first 3D image; obtaining a first 2D image based on the first 3D image; obtaining hand joint information based on the first 2D image; and determining 3D position information of the hand joints based on the first 3D position information and the hand joint information to determine gesture information; in a far-field interaction mode, acquiring a second 3D image containing a hand captured by a 3D camera; acquiring a second 2D image containing a hand captured by a 2D camera; obtaining second 3D position information of the hand based on the second 3D image; obtaining hand joint information based on the second 2D image; and performing coordinate fusion on the second 3D position information and the hand joint information to determine the 3D position information of the hand joints to determine gesture information.
Owner:BEIJING SHIYAN TECH CO LTD
An unmanned aerial vehicle closed-loop flight control system and method based on multi-modal dynamic gesture recognition
This invention discloses a closed-loop flight control system and method for unmanned aerial vehicles (UAVs) based on multimodal dynamic gesture recognition, belonging to the field of UAV flight control. The method includes: acquiring multimodal gesture video streams; extracting spatial and frequency features using a model combining residual networks and FreqFormer, then fusing the extracted gesture prediction categories; quantizing the model using INT8 and deploying it on an airborne NPU based on a producer-consumer model; using a recognition queue to perform mode smoothing on the gesture prediction categories to obtain valid gesture commands, which are then mapped to flight control targets; combining positioning information with cascaded PID control to calculate rotational speed signals, and introducing a first-order inertial link model for delay compensation; improving recognition accuracy in complex environments through multimodal fusion and frequency domain attention mechanisms; ensuring airborne real-time performance through quantization and parallel processing; and significantly enhancing the smoothness and closed-loop stability of UAV gesture control by combining smoothing and delay compensation techniques.
Owner:TIANMUSHAN LABORATORY +1
Electronic device unlocked and controlled by gesture recognition
ActiveUS12646352B2Sensing radiation from moving bodiesDigital data authenticationControl electronicsHuman–computer interaction
There is provided an electronic device capable of improving the difficulty of breaking a digital electronic lock and the accuracy of gesture control. The electronic device is unlocked according to gestures of a single hand or two hands, an operation hand, a gesture position, a staying time of gesture and a gesture variation of a user. The electronic device further improves the accuracy of gesture control in conjunction with an accessory pattern, temperature sensing and space information of a wearable device. The present disclosure further provides an electronic device that identifies legitimacy of moving the electronic device according to a variation sequence of 3-axis accelerations of an accelerometer.
Owner:PIXART IMAGING INC
A gesture recognition-based method for framing a target range
PendingCN122176786AInput/output for user-computer interactionCharacter and pattern recognitionFault toleranceEngineering
This invention relates to the field of target range definition technology, and in particular to a method for target range definition based on gesture recognition. The technical solution includes: target definition and command issuance can be completed through gestures, improving the naturalness and convenience of human-computer interaction; through multi-feature point recognition, perspective correction, anomaly point screening, and edge smoothing, the accuracy and stability of the defined range are improved, enhancing the accuracy and robustness of target definition; the system has automatic and manual adjustment modes, and can autonomously optimize or prompt user intervention in different environments, improving the overall reliability of the system and enhancing its adaptability and fault tolerance; the system can not only define the range, but also analyze multiple targets within the range and automatically select the optimal target to follow based on a preset strategy; through a layered architecture design, gesture recognition, data processing, and motion control are organically combined to ensure that the robot can accurately understand the user's intentions and execute stable and continuous following actions.
Owner:钱钟浩 +1
An interaction design method and system based on gesture recognition
This application provides an interaction design method and system based on gesture recognition. The method includes: determining whether a user has started executing a high-precision selection task based on hand movement and posture information; switching the system's working state to a locked state in response to determining that the user has started executing the high-precision selection task; in the locked state, when the user's hand is detected to form a posture indicating a global mode switching command, the global mode switching command is not executed immediately; in the locked state, when the user's dominant hand is detected to form a posture indicating a global mode switching command, and simultaneously the user's non-dominant hand is detected to perform a preset auxiliary confirmation action, the global mode switching command is executed; in the locked state, local operation commands directly related to the high-precision selection task are prioritized for response. This application can effectively avoid accidental switching of the global mode due to gesture misrecognition, while prioritizing the response to local operation commands, ensuring the smooth execution of high-precision tasks.
Silk fabric sensor and its application in gesture recognition
The application provides a silk fabric sensor and application thereof in gesture recognition, and belongs to the technical field of flexible electronics and wearable sensing. X MXene nanosheets are functional fillers, and Ti X MXene nanosheets are loaded on the surface of silk fibers through an immersion coating process to form a core-shell structure conductive fabric, and Ti X MXene and silk fibroin are stably combined through a hydrogen bond network and electrostatic interaction, the water vapor transmission rate of the sensor is more than 2000 g / m²・day, the sensitivity coefficient GF in the low strain region of 0-18% is 149.31, and the sensitivity coefficient GF in the high strain region of 18-32% is 24.57. X MXene nanosheets are functional fillers, and stable combination of MXene and silk fibers is realized through optimization of a preparation process, so that the application has high sensitivity, good air permeability and excellent fatigue resistance, and a high-precision and low-delay gesture recognition and human-computer interaction system is constructed.
Owner:ZHEJIANG SCI-TECH UNIV
Real-time gesture recognition system based on 16-way imu array and spatiotemporal feature fusion and application
PendingCN122152113AInput/output for user-computer interactionBiological modelsNetwork structureEngineering
The application discloses a kind of real-time gesture recognition systems based on 16-way IMU array and space-time feature fusion, including data acquisition module, data preprocessing module, space-time feature extraction module and real-time identification display module.The IMU array is composed of 16 six-axis inertial measurement units, and 96-dimensional original motion data is continuously sent in fixed format by UDP protocol;Data preprocessing module uses Euclidean alignment and sliding window segmentation technology, window slicing is carried out according to 30 frame length, 15 frame step, and 96-dimensional data is reconstructed into 6×16×30 three-dimensional tensor.The space-time feature extraction module supports ListenNet, DARNET and three kinds of deep network structures based on multi-channel time sequence patch visual Transformer (ViT), realizes the fusion modeling of local structure feature and global time sequence dependence.Has the advantages of strong data structure retention, excellent cross-subject generalization ability, simple deployment, high real-time performance, etc., applicable to wearable interaction, robot control, sign language recognition and other scenes.
Owner:TONGJI UNIV
A non-contact game character control method for rehabilitation training
The application provides a non-contact game character control method for rehabilitation training, and the method comprises the following steps: in a game setting stage, a preset puppet character action is displayed, and a user is prompted to make an arbitrary gesture; a Leap Motion device continuously captures multiple frames of hand images and extracts gesture feature correlation and saves the gesture feature correlation to a gesture feature library; a gesture recognition interface based on the gesture feature library is created, and a gesture feature quantization strategy is configured to establish a mapping relationship between the gesture feature quantization strategy and a current character action; in a game starting stage, the gesture recognition interface is called to perform gesture recognition and gesture feature quantization, and a preset puppet character is controlled to perform a corresponding action according to a gesture feature quantization result and the mapping relationship. The application can automatically complete arbitrary binding between a gesture and a character action through creation and configuration of a gesture feature recognition interface, and can automatically perform quantization processing according to characteristics of the gesture feature to improve the fluency of gesture control, thereby significantly improving the rehabilitation training effect.
Owner:BEIJING UNIV OF TECH
A gait real-time analysis method and system based on visual skeleton feature extraction
The application discloses a gait real-time analysis method and system based on visual skeleton feature extraction, which comprises the following steps: collecting a user walking video through a single camera; extracting human key skeleton points in real time by using a lightweight AI model; calculating anatomical coronal plane, sagittal plane parameters and gait space-time parameters based on the skeleton point coordinates; superimposing the skeleton and parameters on the video in real time and providing voice feedback; starting and stopping through a gesture recognition non-contact control system; automatically storing data and generating a rehabilitation report; the system comprises a USB camera, an embedded processing end and a feedback end with display and audio; the application replaces expensive motion capture equipment with a low-cost single camera, realizes objective quantitative evaluation of gait training, real-time multi-modal feedback and automatic management, significantly reduces medical costs, improves the efficiency and effect of rehabilitation training, and is especially suitable for primary medical institutions.
Owner:TEHLIN PROSTHETIC & ORTHOPAEDIC (S Z) LTD
A gesture recognition method, system, device and computer readable storage medium
PendingCN122176808ABiological modelsBiometric pattern recognitionPattern recognitionSpatial prediction
This application discloses a gesture recognition method, system, device, and computer-readable storage medium, relating to the field of artificial intelligence technology. The method involves training a gesture recognition model using a first training view and a second training view. The gesture recognition model is built based on a spiking neural network model. The method extracts a first spatial feature sequence and a first spatial prediction vector from the first training view generated by the gesture recognition model. It also extracts a second spatial feature sequence and a second spatial prediction vector from the second training view generated by the gesture recognition model. Based on the differences between the first spatial feature sequence and the second spatial prediction vector, and the differences between the second spatial feature sequence and the first spatial prediction vector, a temporal contrast loss value is generated. The gesture recognition model is adjusted based on the temporal contrast loss value to obtain a trained gesture recognition model. A gesture view is then acquired. Finally, the gesture recognition model is applied to recognize the gesture view to obtain the gesture recognition result. This improves the accuracy of gesture recognition.
Owner:HUNAN NORMAL UNIVERSITY
A puncture navigation interaction system and method based on offline voice and gesture recognition fusion
The present application relates to the technical field of robot navigation, in particular to a puncture navigation interaction system and method based on offline voice and gesture recognition fusion, which completes voice collection and recognition locally through an offline voice recognition module, without network transmission, eliminating the risk of patient privacy leakage and the influence of network delay, and ensuring real-time response in the surgical environment; the gesture recognition module adopts a hands-off collection method, so that the surgical operator can input instructions without touching any physical device, strictly meeting the sterile operation specification; the modal fusion decision unit performs consistency verification and priority scheduling on the dual-modal recognition results, outputs a single effective control instruction, effectively avoids instruction conflicts and misoperations, and significantly improves the interaction fault tolerance; the control unit converts the instruction into a driving signal and controls the puncture navigation robot to perform the corresponding action, realizing non-contact, high-reliability and low-delay human-computer interaction, and comprehensively guaranteeing the safety, accuracy and sterility of the puncture navigation surgery.
Owner:SHANGHAI SIMPLETOUCH ROBOT CO LTD