Gesture Recognition Patch Selection for Long-Distance Control
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing gesture recognition systems struggle to accurately recognize gestures from a distance greater than one meter due to small spatial regions in their field-of-view and limited adaptability to variations among users.
Innovation Solution
A computer-implemented method using a machine learning model with a first subnetwork for feature extraction, a second subnetwork for patch selection, and a third subnetwork for gesture classification, enabling accurate gesture recognition and control of actuators based on sensor data.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If gesture recognition systems use traditional sensor setups, then they can recognize gestures accurately at close range, but they fail to recognize gestures accurately when the subject is more than one meter away
Solution Approach 1:
The patent divides the feature data into a set of patches and selects a target patch from among the patches. This segmentation allows the system to focus on specific regions of interest from the sensor data, enabling accurate gesture recognition at longer distances by processing only the relevant portions of the image rather than the entire field of view.
2Reliability
If gesture recognition systems are trained to recognize specific gestures, then they can identify known gestures, but they cannot adapt to variations of gestures performed by different users
Solution Approach 1:
The patent employs a machine learning model that dynamically adapts to different users and gesture variations. The model includes a first subnetwork for feature extraction, a second subnetwork for patch selection, and a third subnetwork for gesture classification, enabling the system to learn and adapt to variations in gesture performance across different users without requiring complete retraining.
Data Source
AI summary
A computer-implemented system and method relate to gesture recognition. A machine learning model includes a first subnetwork, a second subnetwork, and a third subnetwork. The first subnetwork generates feature data based on sensor data, which includes a gesture. The feature data is divided into a set of patches. The second subnetwork selects a target patch of feature data from among the set of patches. The third subnetwork generates gesture data based on the target patch of feature data. The gesture data identifies the gesture of the sensor data. Command data is generated based on the gesture data. A device is controlled based on the command data.