Computed Tomography Image Analysis of Body Fat Based on Multi-Image Information

Area of Science:

• Medical Imaging
• Computer Vision
• Biomedical Engineering

Background:

• Objective health assessment necessitates body fat measurement in both obese and nonobese individuals.
• Image-based body fat analysis offers potential for accelerated diagnostic capabilities.
• Incorporating age and sex into body fat analysis can enhance disease correlation and fat distribution insights.

Purpose of the Study:

• To evaluate existing computed tomography (CT) imaging algorithms for abdominal and subcutaneous fat identification and segmentation.
• To present an enhanced region growing scale-invariant feature transform (SIFT) algorithm for improved body fat analysis.
• To optimize image processing for rapid and accurate comparison of images across multiple databases.

Main Methods:

• Evaluation of CT imaging algorithms for human abdominal and subcutaneous fat segmentation.
• Development of an improved region growing SIFT algorithm incorporating Naive Bayes image thresholding.
• Implementation of key point selection and image matching techniques for enhanced accuracy.

Main Results:

• The proposed algorithm demonstrates improved efficiency in image processing for body fat analysis.
• The method enables rapid and accurate comparison and matching of images from diverse databases.
• Enhanced segmentation and identification of abdominal and subcutaneous fat were achieved.

Conclusions:

• The improved region growing SIFT algorithm offers a more efficient and accurate approach to image-based body fat assessment.
• This advancement supports faster diagnosis and better correlation of diseases with fat distribution.
• The developed image processing techniques are valuable for objective health evaluations.