Gesture Key Point Detection via Category-Specific Models
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current gesture recognition technologies face challenges in accurately detecting key points from images, as they often rely on general models that do not account for specific gesture categories, leading to reduced detection accuracy.
Innovation Solution
A method that involves acquiring a target image, determining the gesture category, and selecting a corresponding key point detection model from a set of models trained for each gesture category, allowing for precise key point detection by using parameters tailored to the specific gesture category.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If a general key point detection model is used for all gesture categories, then the device complexity is reduced, but the measurement precision of gesture key points deteriorates
Solution Approach 1:
The patent divides the gesture recognition task into multiple segments by creating separate key point detection models for different gesture categories (e.g., hand gestures, foot gestures, different gesture types). Each model is specialized for its specific gesture category, which improves detection accuracy for each segment while managing overall system complexity through modular organization.
Solution Approach 2:
The patent applies local quality by tailoring the detection model parameters and structures to match specific gesture categories. Each key point detection model has locally optimized parameters (such as key point locations, detection thresholds, and model architectures) that are specifically suited for its designated gesture category, thereby improving overall measurement precision.
2Measurement precision
If multiple key point detection models are maintained for different gesture categories, then the measurement precision improves, but the loss of time for model selection and management increases
Solution Approach 1:
The patent implements preliminary action by pre-categorizing gestures and pre-configuring appropriate key point detection models for each gesture category before actual detection occurs. The system预先 establishes the mapping between gesture categories and their corresponding optimized models, so that during runtime, only quick category identification and model selection are needed, minimizing time loss.
Solution Approach 2:
The system employs self-service mechanisms where the gesture category automatically selects the appropriate key point detection model without requiring manual intervention. The system autonomously manages the multiple models by automatically identifying which model to use based on the detected gesture category, reducing the time and effort required for model management.
3Adaptability or versatility
If gesture category determination is performed before key point detection, then the adaptability of the detection system improves, but the duration of action of the detection process increases
Solution Approach 1:
The detection process is segmented into two distinct phases: gesture category determination and key point detection. By separating these tasks, the system can efficiently handle different gesture types through specialized models while maintaining overall adaptability. The segmentation allows for optimized processing in each phase without significantly increasing total detection time.
Solution Approach 2:
The system utilizes parameter changes by adjusting detection parameters (such as model selection, key point coordinates, and detection thresholds) based on the determined gesture category. This dynamic parameter adjustment enables the system to adapt to different gesture categories while using pre-configured models, thereby minimizing the additional time required for adaptation.
Data Source
AI summary
A method for detecting gesture key points can include: acquiring a target image to be detected; determining a gesture category according to the target image, the gesture category being a category of a gesture contained in the target image; determining a target key point detection model corresponding to the gesture category from a plurality of key point detection models; and performing a key point detection on the target image by the target key point detection model.