Multi-Camera Hand Motion Detection Using Wrist-Guided Image Fusion
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing systems struggle to accurately detect the motion of a user's hand due to limitations in camera angle of view, which prevents the capture of both the user's body and hand, leading to incomplete image analysis and reduced detection accuracy.
Innovation Solution
The system employs multiple cameras positioned to capture images of the user's body and wrist separately, combining these images to generate a composite view that allows for the detection of feature points and motion of the hand using projective and optical-axis transformations.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Device complexity
If a single camera is used to capture the user's body and hand, then the device complexity is reduced, but the measurement precision of hand motion detection deteriorates due to insufficient coverage
Solution Approach 1:
The system divides the imaging task into multiple segments by using separate cameras for different body parts. The first camera captures the user's body while the second camera captures the hand and wrist, allowing each camera to be optimized for its specific target and ensuring complete coverage without requiring a single complex high-angle camera
Solution Approach 2:
The system transitions from a single-viewpoint 2D image capture to a multi-viewpoint 3D spatial reconstruction. By capturing images from different camera positions and combining them through coordinate transformation, the system creates a comprehensive spatial model that improves measurement precision without proportionally increasing device complexity
2Measurement precision
If the camera angle of view is increased to capture both body and hand, then the measurement precision improves, but the device complexity increases due to additional cameras and processing
Solution Approach 1:
Rather than using one camera with an excessively wide angle that would compromise image quality, the system segments the imaging function across multiple cameras with appropriate individual fields of view. This allows each camera to capture its target area with optimal precision while the combined system achieves comprehensive coverage
Solution Approach 2:
The system introduces coordinate transformation and image synthesis algorithms as intermediaries that bridge the gap between multiple camera views. These processing steps combine the data from simpler individual cameras to produce the equivalent information that would require a single complex camera system
Data Source
AI summary
In an information processing device, a processor acquires a first image including a body of a user captured by a first camera. The processor acquires a second image including a wrist of the user captured by a second camera. The processor generates a third image including the body and the wrist of the user by combining the first image and the second image. The processor determines a plurality of first feature points corresponding to the body of the user from the third image. The processor acquires a fourth image including a hand of the user based on a third feature point corresponding to the wrist of the user. The processor determines a plurality of second feature points corresponding to the hand of the user from the fourth image. The processor detects a motion of the hand of the user based on the plurality of second feature points.