Joint Center Detection via Iterative Visual Hull Segmentation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing 3D motion capture systems face challenges in accurately and reliably segmenting joint center locations due to low contrast, image noise, and occlusions, which hinders clinical applications such as longitudinal tracking and database normative referencing.
Innovation Solution
The system employs advanced image processing techniques, including iteratively trained deep neural network algorithms, to detect joint centers by capturing digital images, applying visual hull segmentation, and updating segmentation models based on functional movements, enabling accurate identification of various joint types in the human body.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If traditional segmentation algorithms are used, then the system is simpler to implement, but the accuracy and reliability of joint center detection deteriorates due to low contrast, image noise, and occlusions
Solution Approach 1:
The system performs preliminary actions by capturing functional movement data before final segmentation. The method captures digital images showing functional movement of the human subject, then uses this movement information to guide the segmentation process. This preliminary capture of movement data enables the system to distinguish anatomical structures more reliably even in challenging imaging conditions.
Solution Approach 2:
The system implements feedback by using detected joint center locations and segmentation results to refine and update the segmentation model iteratively. The method updates the segmentation model based on the detected joint centers, creating a feedback loop that continuously improves detection accuracy. This iterative refinement process allows the system to overcome initial segmentation errors caused by noise and low contrast.
2Adaptability or versatility
If a general body segmentation method is developed, then the adaptability to various applications increases, but the reliability and repeatability of joint center location deteriorates due to imaging ambiguities
Solution Approach 1:
The system applies local quality by using context-specific segmentation algorithms tailored to different body regions and joint types. The method selects segmentation algorithms based on the context of the visual hull segmentation, applying different algorithms for different anatomical structures. This localized approach ensures optimal performance for each specific application while maintaining overall system reliability.
Solution Approach 2:
The system utilizes parameter changes by adjusting segmentation parameters based on functional movement characteristics. The method applies deep neural network algorithms that utilize functional movement data to select and configure segmentation parameters dynamically. This allows the system to adapt to different imaging conditions and anatomical variations while maintaining consistent and reliable joint center detection across diverse applications.
3Measurement precision
If visual hull segmentation with proxy spheres is used, then body segments can be represented, but the precision of joint center coordinates deteriorates without iterative refinement
Solution Approach 1:
The system performs preliminary action by using visual hull segmentation with proxy spheres to establish initial joint center coordinates before refinement. This initial segmentation provides a reasonable starting point that reduces the computational burden of subsequent refinement steps, achieving a balance between speed and precision.
Solution Approach 2:
The system maintains continuity of useful action by implementing an iterative refinement process that continuously improves joint center coordinate precision. The method updates the segmentation model based on detected joint centers and functional movement data, performing multiple refinement cycles until convergence. This continuous refinement ensures high precision without excessive computational overhead by stopping when improvement becomes negligible.
Data Source
AI summary
The disclosure is related to methods and systems for digital image segmentation for objectively identifying joint center locations. In one embodiment, the methods and system disclosed herein use automated methods to capture and process digital images and motion data to identify, validate, and apply segmentation algorithms trained for one or more targeted anatomical structures of a human subject.