Gesture Identification Using Frame Subtraction and Motion Vectors
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current gesture-control systems using natural images without auxiliary illuminants face challenges in identifying gestures due to high error rates, complexity, and cost, particularly with Z-axis motions and indistinct images, limiting their applicability to short distances and requiring predefined shapes.
Innovation Solution
A method that generates variant images by subtracting successive frames, extracts image features like barycenter or variance, and compares these patterns with predefined gestures, eliminating the need for shape recognition and position detection, thus being insensitive to image quality and color temperature, and supporting X, Y, and Z-axis motions without fixed gestures.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If shape detection method is used for gesture identification, then the system can recognize predefined gestures, but it requires high-definition images and cannot adapt to indistinct images caused by fast motions
Solution Approach 1:
The patent changes the parameter being measured from spatial shape characteristics to temporal motion characteristics. By computing motion vectors between consecutive frames and analyzing motion patterns, the system becomes insensitive to image quality while maintaining gesture recognition accuracy. This parameter transformation resolves the contradiction between requiring high-definition images and adapting to indistinct images.
2Adaptability or versatility
If motion estimation method is used for gesture identification, then the system can identify gestures with Z-axis motions, but it is weak in identifying certain gestures due to user-specific operational habits
Solution Approach 1:
The patent creates a universal gesture recognition system that handles multiple gesture types (including Z-axis motions, horizontal waves, vertical waves, and clicks) through a single motion-based framework. By analyzing motion vectors and patterns rather than relying on user-specific shape characteristics, the system achieves both versatility in supporting various gesture dimensions and accuracy in identification.
3Measurement precision
If skin-color analysis and shape recognition are used, then the system can identify hand regions, but it requires complex algorithms and numerous arithmetic operations
Solution Approach 1:
The patent extracts only the essential motion information from the image sequence by computing motion vectors between consecutive frames. This extraction approach eliminates the need for complex skin-color analysis and shape recognition algorithms, reducing computational complexity while maintaining the ability to identify hand regions and gestures through motion patterns alone.
4Measurement precision
If conventional shape detection method is used, then the system can recognize fixed gestures, but it greatly restricts the scope of operational gestures
Solution Approach 1:
The patent transitions from static shape-based gesture recognition to dynamic motion-based recognition. By analyzing motion vectors and temporal patterns of hand movement, the system can recognize a diverse range of gestures including those with Z-axis motions, horizontal waves, vertical waves, and clicks, greatly expanding the scope of operational gestures beyond fixed predefined shapes.
Data Source
AI summary
A method for gesture identification with natural images includes generating a series of variant images by using each two or more successive ones of the natural images, extracting an image feature from each of the variant images, and comparing the varying pattern of the image feature with a gesture definition to identify a gesture. The method is inherently insensitive to indistinctness of images, and supports the motion estimation in axes X, Y, and Z without requiring the detected object to maintain a fixed gesture.