23 results about "Gesture analysis" patented technology

An optical sensing system comprises a human interface structure to be applied to a user. The system comprises a time-of-flight sensor with an array of laser diodes with Bragg reflectors operating in the near-infrared, a detection system, and a processor generating a time-of-flight measurement and a first image. The system also includes a first camera system capturing a second image, and a second camera system illuminated by light emitting diodes to monitor one or more eyes of the user. In communication with or operating on the processor is an artificial intelligence assistant, which comprises a multi-modal generative artificial intelligence model. The system may be part of wearable eyewear such as smart glasses or extended reality goggles. The system may perform gesture analysis, eye tracking and speed recognition, and video and verbal outputs can be provided to the user. An artificial intelligence agent may have awareness of the environment.

The invention provides a touch screen input method and system, and relates to the technical field of touch recognition of electronic equipment.The touch screen input system comprises the steps that an input completion instruction of a user to a touch screen in each application scene is obtained; according to the input completion instruction, collecting an operation gesture corresponding to completion of the input completion instruction; according to the operation gesture and the application scene thereof, establishing operation sets corresponding to different application scenes; performing gesture analysis on each operation set to obtain a habitual gesture model; detecting a target gesture of the user through the habit gesture model to obtain a detection result; and according to the detection result, matching a corresponding target input completion instruction to complete instruction input. According to the input method and system provided by the invention, the input completion instruction response timeliness and accuracy can be ensured when the user inputs the operation gesture on the touch screen, so that the operation gesture input process is more intelligent.

Systems and methods for displaying augmented reality clinical movements can use an augmented reality device to display various aspects of clinical movements. The systems and methods can use a motion capture device to capture clinical movements. A method can include analyzing information about a clinical movement to determine a motion path representing at least a portion of the clinical movement. The method can automatically define a path area or a virtual target in an augmented reality environment overlaid on a real environment. The method can display the path area or the virtual target on the augmented reality display.

The application is suitable for the technical field of lighting control, and provides a lighting device control method and device based on wireless communication and a lighting device, the method comprising: establishing a wireless communication connection with a control terminal, and periodically acquiring a received signal strength indication value; performing sliding window filtering processing on the value, and when the value rises and breaks through a first preset threshold, waking up the lighting device to enter a gesture analysis window period and starting a timer; continuously recording a change track of the received signal strength indication value in the gesture analysis window period, and generating a received signal strength time sequence; extracting extreme value features in the received signal strength time sequence, and matching the extracted extreme value features with a preset signal action template to determine a corresponding control instruction; and according to the matched control instruction, directly executing a corresponding lighting state switching operation locally by the lighting device. The application solves the problem that it is difficult to quickly and effectively operate the lighting device in the existing industrial and mining environment.

The application discloses a motion posture analysis method, earphones, an earphone system and a computer readable storage medium. The posture analysis method is applied to the earphones. The method comprises the following steps: acquiring a motion signal, wherein the motion signal comprises an acceleration signal; calculating a running posture parameter based on the acceleration signal, wherein the running posture parameter comprises at least one of a vertical amplitude and a ground contact time, the vertical amplitude represents a vertical displacement variation of the earphones relative to a horizontal plane, and the ground contact time represents a time when a single foot of a user contacts the horizontal plane; and generating prompt information based on the running posture parameter. The vertical amplitude reflects the work condition of the user in the vertical direction, and monitoring the vertical amplitude can guide the user to reduce unnecessary vertical jumping and improve energy utilization efficiency. The ground contact time quantifies the duration of the contact between the foot of the user and the horizontal plane, thereby providing a basis for evaluating the running efficiency and the motion impact force. In this way, the motion posture obtained based on the vertical amplitude and the ground contact time parameter has high accuracy, and the generated prompt information is highly professional.

The application discloses a kind of based on gesture analysis's radioactive source equipment abnormal transaction detection method and system, specifically related to the measurement field of radiation beam current, including obtaining the attitude data of radioactive source equipment, by analyzing attitude data and constructing beam principal axis vector to express the radiation propagation direction of radioactive source equipment under current attitude;By constructing radiation action region and judging its space inclusion relation to identify whether the irradiation area of radioactive source under current attitude is out of bounds;By analyzing the detection data to construct the measured intensity principal axis vector and compare with beam principal axis vector to identify whether the radiation propagation direction is shifted. By constructing the mapping relationship between device attitude and beam principal axis, and jointly judging the out-of-bounds state of space action region and the shift state of radiation propagation direction, the accurate identification of posture-induced radiation directionality abnormality is realized, so as to effectively solve the problem that the traditional system cannot perceive beam out-of-control path.

The invention belongs to the technical field of holographic interaction and edge computing, and discloses an off-line multi-mode holographic sand table interaction method and system based on edge computing. The problems that existing holographic sand table interaction depends on the cloud, multi-mode anaphora disambiguation is difficult, delay is high, and safety is insufficient are solved. The method runs offline in the whole process, voice and three-dimensional gesture streams are collected, and time-space synchronization and coordinate unification are completed; multi-modal features are analyzed in parallel through a lightweight model, and three-dimensional matching fusion is achieved through an annular buffer queue and spatio-temporal correlation scores; true three-dimensional holographic display is realized based on voxel cloud rendering, continuous interaction logic is established in combination with a finite state machine, and rendering information is transmitted back to gesture analysis to form closed-loop feedback. The system comprises an acquisition module, an analysis module, a fusion module, a rendering module and a holographic display terminal, can locally realize low-delay and high-precision multi-mode interaction, is suitable for secret-related and network-free command decision-making scenes, and improves interaction naturalness and data security.

The application discloses a kind of vascular intervention operation action sensing and learning system, method and computer storage medium, including operation gesture acquisition unit, operation gesture analysis unit, intervention instrument accurate delivery unit;The operation gesture of vascular intervention operation of operator is sensed using multiple sensors, data cleaning is realized by Gaussian clustering, the learning of machine on time sequence to doctor gesture and corresponding effect is realized using deep learning model CNN-LSTM, and the autonomous operation of vascular intervention robot is realized according to relevant prediction results.The operation action of person can be converted to the action of machine autonomous operation in the application, and the machine has corresponding action learning function.Through the autonomous learning of surgical robot operation, the surgical robot gradually has the ability of independent work, and autonomously executes surgical tasks, so as to reduce various surgical risks caused by human factors.

An early warning method for tumbling during riding, comprising: conducting riding monitoring to a target rider, and obtaining a first riding velocity, conducting riding monitoring to the target rider, and obtaining a second riding velocity and acceleration data; where the communication status of the GNSS positioning module is regular, monitoring the first riding velocity; where the acceleration data satisfy an interval of a first threshold, giving threshold analysis for the first riding velocity and giving riding gesture analysis to the target rider and obtaining riding gesture information; where the first riding velocity satisfy an interval of a second threshold, and the riding gesture data show irregular riding gestures, pronouncing a warning to the target rider, collecting operation data of the target rider, where the operation data of the target rider are not collected, sending assistance request data, and sending the assistance request data to riders in a target area.

A three-dimensional gesture detection device and a three-dimensional gesture detection method. The three-dimensional gesture detection device includes a node detection unit, a gesture recognition model, and a gesture trajectory detection unit. The node detection unit obtains a plurality of nodes according to each hand image of a continuous hand image. The gesture recognition model obtains a plurality of gesture class confidences. The gesture trajectory detection unit includes a weight analyzer, a gesture analyzer, a key point classifier, and a trajectory analyzer. The weight analyzer obtains a plurality of weights of the plurality of gesture classes according to a user interface. The gesture analyzer performs weighted calculation on the plurality of gesture class confidences to analyze a gesture of each image. The key point classifier classifies a plurality of key points from the plurality of nodes. The trajectory analyzer obtains an inertia trajectory of a gesture according to the plurality of key points.

The invention relates to the technical field of rail transit visual inspection, and discloses a subway driver standardized operation behavior recognition cascade early warning system and method for edge calculation. A multi-small-model cascade architecture is adopted, behavior recognition is realized through Yolo11 + Transform, gesture analysis is realized through OpenPose + ST-GCN + SoftMax, and in combination with dynamically quantized model lightweight processing and a multi-model fusion strategy, the reasoning speed is greatly increased and the computing power dependence is reduced on the premise of ensuring the recognition accuracy. According to the method, cloud processing is not needed, real-time monitoring and early warning can be independently completed on edge equipment, full-dimension supervision from macroscopic behaviors to microscopic gestures is achieved, the problems that in the prior art, models are complex, the real-time performance is poor, and the service fitting degree is insufficient are effectively solved, and reliable guarantee is provided for safe operation of subways.

Methods, systems, and apparatus, including computer programs encoded on computer storage media, for predicting a user's engagement. In some implementations, a system obtains data indicative of a time evolving movement of interactions of a user with a website shown on the client device. The system determines, using a first trained machine learning model and based on the data indicative of the time evolving movement of the user, a metric associated with an emotion of the user corresponding to the user's interaction with the website. The system obtains, from a metric database, one or more metrics associated with an identifier of the user. The system provides, to a second trained machine learning model, (i) the metric associated with the emotion and (ii) data representing the obtained metrics. In response, the system generates, using the second trained machine learning model, a prediction indicating an engagement level of the user.

An early warning method for tumbling during riding, comprising: conducting riding monitoring to a target rider, and obtaining a first riding velocity, conducting riding monitoring to the target rider, and obtaining a second riding velocity and acceleration data; where the communication status of the GNSS positioning module is regular, monitoring the first riding velocity; where the acceleration data satisfy an interval of a first threshold, giving threshold analysis for the first riding velocity and giving riding gesture analysis to the target rider and obtaining riding gesture information; where the first riding velocity satisfy an interval of a second threshold, and the riding gesture data show irregular riding gestures, pronouncing a warning to the target rider, collecting operation data of the target rider, where the operation data of the target rider are not collected, sending assistance request data, and sending the assistance request data to riders in a target area.

A system comprising: a control device and several interaction devices configured to communicate wirelessly with the control device, each of the several interaction devices comprising several sensory stimulation elements, a motion sensor, at least one wireless communication device and a user input element, and the several sensory stimulation elements comprising one or more light sources, audio sources and vibration sources; wherein the control device can be communicatively connected to an activity routine database comprising several activity routines;and wherein the control device and the multiple interaction devices are configured to implement a method of interactive physical and cognitive training based on multisensory external stimulation and body gesture detection, comprising the following steps: (B) providing an activity routine database comprising multiple activity routines, wherein the control device is communicatively coupled to the activity routine database; (C) communicatively coupling the control device with at least one primary interaction device from among the multiple interaction devices; (D) communicatively coupling each of the at least one primary interaction device with at least one secondary interaction device from among the multiple interaction devices by means of the at least one wireless communication device;(E) Receiving an activity routine selection by the control device, wherein the activity routine selection corresponds to a specific activity routine from the multiple activity routines; (F) Actuating the sensory stimulation elements of the multiple interaction devices according to the specific activity routine; (G) Receiving a series of user interaction signals with the control device through the user input elements of the multiple interaction devices; (H) Monitoring multiple activity data received by the control device from the motion sensor of a specific interaction device from the multiple interaction devices; (I) Analyzing the activity data by the control device using a gesture analysis algorithm; (J) Actuating at least one of the sensory stimulation elements of the specific interaction device based on the analysis of the multiple activity data;and (K) the control device analyzes the series of user interaction signals with an activity analysis algorithm to generate multiple activity performance data; the system is further configured to: receive a benchmark routine selection as the activity routine selection by the control device, wherein the benchmark routine selection corresponds to a benchmark routine from the multiple activity routines and wherein the benchmark routine is associated with multiple performance metrics; actuate the sensory stimulation elements of the multiple interaction devices according to the benchmark routine; analyze the series of user interaction signals with the multiple performance metrics to generate multiple benchmark data as the multiple activity performance data by the control device; compare the multiple benchmark data with the multiple performance metrics by the control device;Generating at least one personalized activity recommendation based on comparison by the control device; and storing the at least one personalized activity recommendation as one of several activity routines in the activity routine database by the control device.

The application provides a standby picture processing method of a Micro LED display screen, comprising: if the track overlap analysis result shows that a plurality of touch point track overlaps, separating the overlapping tracks through a clustering algorithm, calculating track overlap degree and separation confidence, and obtaining a touch track separation result; combining the pressure mutation detection result with the touch point coordinate change, calculating the touch offset through weighted average, obtaining the touch offset correction parameter, adjusting the touch coordinate mapping relationship according to the correction parameter; through the touch offset correction result, combining the user finger contact angle and contact area change, identifying the user touch gesture mode by adopting a gesture recognition technology, extracting the finger gesture feature parameter, and generating the user touch gesture data; according to the touch offset correction result and the pressure mutation detection result, combining the finger gesture feature parameter, modeling the user individual touch habit through gesture analysis, and obtaining the touch sensitivity adjustment parameter.

Methods, systems, and apparatus, including computer programs encoded on computer storage media, for predicting a user's receptivity. In some implementations, a system obtains data indicative of a time evolving movement of a user with a website shown on the client device. The system determines, using a first trained machine learning model, a likelihood for each of a set of emotions of the user associated with the interactions of the user with the website. The system extracts, from the website, a plurality of visible features from the website. The system provides, to a second trained machine learning model, (i) the determined likelihood for each of the set of emotions and (ii) data that represents the plurality of visible features. In response, the system generates, using the second trained machine learning model, a prediction indicating a receptivity score of the user associated with the website.

The invention provides a VR game scene real-time rendering method based on 5G. The method specifically comprises the steps that a type identifier of each data packet in a network transmission layer is acquired, whether the data packet is a key data packet or a secondary data packet is judged through a preset threshold value, and a priority distribution table is obtained; dynamically adjusting a network transmission queue in a high-concurrency environment according to the priority distribution table, if the data packet identifier is a key type, immediately improving the transmission sequence, and determining the adjusted transmission sequence; extracting a key data packet from the adjusted transmission sequence, transmitting the key data packet to a calculation processing layer, and matching a corresponding gesture analysis model through a pre-established mapping rule to obtain initial aiming trajectory data; and obtaining the synchronously optimized frame sequence, updating the picture content through a refreshing mechanism of the head-mounted display equipment, and determining final picture output data.

The invention discloses a motion posture analysis system based on artificial intelligence and image recognition. The motion posture analysis system comprises a common RGB camera, a depth camera and a high-speed camera. According to the motion posture analysis system based on artificial intelligence and image recognition, the advantages of a common RGB camera, a depth camera and a high-speed camera are fused through a multi-view camera array, so that two-dimensional and three-dimensional and static and dynamic all-dimensional data acquisition is realized; the problem of data missing caused by limb shielding in single-person movement and personnel overlapping in multi-person movement is effectively avoided, the image quality is optimized through combination of Gaussian filtering and histogram equalization, simultaneous detection of multiple persons is achieved through algorithm fusion of OpenPose and HRNet, the key point positioning precision is improved through high-resolution feature map fusion, and the method is suitable for large-scale popularization and application. The system simultaneously covers static attitude analysis and dynamic attitude analysis, and adopts attitude modeling to quantify an action standard, so that the objectivity of motion attitude analysis is improved, the quantitative standard is clarified, and the analysis efficiency is improved.

The invention relates to the technical field of image processing, in particular to a deep learning-based facial expression and posture analysis attention assessment method, which comprises the following steps of: acquiring a head posture angle sequence of a historical inspector, and generating standard stage maps in sections according to operation actions; acquiring inertial data and visual odometer data of monitoring equipment, and constructing motion state estimation; performing correlation compensation on the head posture angle sequence and motion state estimation to obtain a pure head posture component; extracting eye circumference area key points under face shielding, and calculating the distance change of the eye circumference key points to obtain a tensity index; matching the pure head posture component with a standard stage atlas, and identifying a current operation stage for grading judgment: if in a target approaching or moving transition stage, judging according to the deviation degree of the pure head posture component and a standard track; if in the bending examination or getting-up marking stage, when the posture deviation degree exceeds a threshold value and the tensity index is attenuated, it is judged that an attention distraction state is entered, and early warning is conducted in advance.

The invention relates to the technical field of vision testing, in particular to a dynamic sighting mark generation and gesture feedback interactive vision testing system which comprises a sighting mark generation module, a gesture recognition module and a feedback adjustment module. The sighting mark generation module generates an initial sighting mark according to a preset rule and dynamically adjusts the size, position and shape of the initial sighting mark; the gesture recognition module captures gesture actions through a camera and analyzes direction, speed and force information; and the feedback adjustment module calculates sighting mark adjustment parameters based on the gesture analysis result, and optimizes a display strategy in combination with ambient light and screen brightness. Through non-contact gesture interaction and dynamic sighting mark adjustment, the accuracy and the individuation degree of the test result are improved, predictive adjustment is achieved based on deep learning, and the user experience and the intelligent level are remarkably improved.

The invention provides a standby picture processing method for a Micro LED display screen, and the method comprises the steps: separating overlapped tracks through a clustering algorithm if a track overlapping analysis result shows that the tracks of a plurality of touch points are overlapped, calculating the track overlapping degree and the separation confidence coefficient, and obtaining a touch track separation result; combining the pressure mutation detection result with the coordinate change of the touch point, calculating a touch offset through weighted average to obtain a touch offset correction parameter, and adjusting a touch coordinate mapping relation according to the correction parameter; recognizing a user touch posture mode by adopting a posture recognition technology through a touch offset correction result in combination with the change of the finger contact angle and the contact area of the user, extracting finger posture feature parameters, and generating user touch posture data; and according to the touch offset correction result and the pressure mutation detection result, combining with the finger posture characteristic parameters, modeling personalized touch habits of the user through posture analysis, and obtaining touch sensitivity adjustment parameters.

The invention provides a camera implementation method and system based on dual-camera data fusion and a storage medium, and the method comprises the steps: firstly, synchronously obtaining face image data and gesture image data of a user through a front camera and a rear camera, and carrying out the timestamp alignment processing, and forming dual-path data streams with time synchronization; then, inputting the face data into a pre-trained emotion recognition model to output an emotion type and a confidence coefficient thereof, and inputting the gesture data into a gesture analysis model to output a gesture type and a confidence coefficient thereof; thirdly, on the premise of time alignment, confidence screening and fusion verification are carried out on the two recognition results, and an effective fusion event is generated; and finally, matching the event based on a preset semantic rule base, and mapping to generate a corresponding camera interaction instruction. According to the method, more natural and intelligent man-machine interaction is realized through multi-dimensional data fusion and collaborative decision of emotions and gestures, and intention understanding accuracy and operation fluency of a camera in a complex scene are improved.

To minimize unintended movement of medical imaging equipment by the operator. [Solution] The medical image diagnostic apparatus according to the embodiment comprises a first sensor unit, a first identification unit, a second sensor unit, a reading unit, an analysis unit, and a control unit. The first sensor unit detects contact by the operator. The first identification unit identifies the object to be operated on. The second sensor unit detects the operator's gestures. The reading unit reads the gestures based on the detection results of the second sensor unit. The analysis unit analyzes the instructions based on the reading results from the second sensor unit. The control unit controls the movement of the object to be operated on, which has been identified by the first identification unit, in a direction corresponding to the analysis results of the analysis unit.