On segmenting the three-dimensional scan data of a human body

Area of Science:

• Computer Vision
• Biomedical Engineering
• Human-Computer Interaction

Background:

• Accurate human anthropometric landmark identification is crucial for various applications.
• Existing methods for processing 3-D surface data can be complex and time-consuming.
• A need exists for robust and efficient software to segment and analyze large 3-D point clouds.

Purpose of the Study:

• To present novel software for categorizing large clouds of 3-D surface data from human subjects.
• To describe an incremental approach for progressively refining the identification of human anthropometric landmarks.
• To detail a step-by-step method for orienting and segmenting 3-D body scan data.

Main Methods:

• Development of specialized software to process over 300,000 3-D surface data points.
• Implementation of an incremental approach for landmark identification refinement.
• Utilization of a discrete point cusp detector algorithm for data cloud separation.

Main Results:

• The software successfully orients and segments human body data points.
• The discrete point cusp detector algorithm effectively separates touching cylindrical objects in 3-D data.
• The software demonstrated robustness across over one hundred diverse body scan datasets.

Conclusions:

• The presented software offers a robust solution for analyzing 3-D human body scan data.
• The discrete point cusp detector is a key algorithm for accurate data segmentation.
• This approach advances the automated identification of human anthropometric landmarks.