A Portable Stereo Vision System for Whole Body Surface Imaging.

Area of Science:

• Computer Vision
• 3D Imaging
• Biomedical Engineering

Background:

• Accurate 3D body surface reconstruction is crucial for various applications.
• Existing systems often face challenges with cost, complexity, or accuracy.

Purpose of the Study:

• To develop a compact, economical, and accurate whole body surface imaging system.
• To introduce a novel stereo matching algorithm for precise body segmentation and geometric detail recovery.

Main Methods:

• Utilized a four-stereo-unit configuration for full body imaging.
• Developed a two-phase sub-pixel, dense stereo matching algorithm.
• Employed block matching for initial disparity map generation and local least squares with global optimization for refinement.

Main Results:

• The system successfully segmented the body from the background.
• Achieved high accuracy in recovering geometric details of the body surface.
• Experimental results demonstrated realistic capture of smooth and natural whole body shapes.

Conclusions:

• The proposed stereo vision system offers a practical solution for whole body surface imaging.
• The novel stereo matching algorithm enhances accuracy and reliability in 3D shape reconstruction.
• This technology has potential applications in fields requiring precise body shape analysis.