Gesture Recognition Device Using Fingertip Candidate Detection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional gesture recognition technologies face challenges in accurately recognizing various gestures in real-time due to limited versatility and increased resource requirements, with existing methods either specializing in specific hand shapes or relying on numerous patterns that enhance accuracy at the cost of resource usage.
Innovation Solution
A gesture recognition device that includes a fingertip candidate detector, a finger detector, and a gesture type specifying unit, which detects fingertip candidates and skin areas extending from them to accurately recognize gestures based on the number of fingers, enabling simple and versatile gesture recognition.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If template matching with multiple patterns is used to enhance gesture recognition accuracy, then recognition accuracy is improved, but resource requirements and processing complexity increase
Solution Approach 1:
The patent extracts only the essential feature of finger count from the complex hand gesture, ignoring other details such as hand shape, finger positioning, and hand orientation. This selective extraction allows the system to achieve high recognition accuracy for the target gesture type while using minimal processing resources and simple algorithms.
Solution Approach 2:
The patent focuses the gesture recognition on a specific local feature (number of fingers) rather than analyzing the entire hand structure. By concentrating on this key local characteristic, the system achieves accurate gesture classification without the need for complex global analysis of the whole hand pose.
2Measurement precision
If specialization in a predetermined hand shape is used, then recognition accuracy for that specific gesture is improved, but versatility for other gesture shapes is reduced
Solution Approach 1:
The patent creates a universal gesture recognition approach by identifying gestures based on the number of fingers rather than specific hand shapes. This single feature-based method can accurately recognize multiple gesture types (one-finger gesture, two-finger gesture, three-finger gesture, four-finger gesture, five-finger gesture) without requiring separate specialized algorithms for each gesture type.
Data Source
AI summary
A gesture recognition device configured to recognize a gesture of a hand from a captured image of a user, having a fingertip candidate detector that detects a fingertip candidate from the image, a finger detector that detects a skin area as a finger, the skin area extending by a predetermined length in a certain direction from the fingertip candidate detected by the fingertip candidate detector, and a gesture type specifying unit that specifies a type of the gesture based on the number of fingers detected by the finger detector.