Binarization of medical images based on the recursive application of mean shift filtering : Another algorithm

Area of Science:

• Medical Image Analysis
• Computer Vision
• Digital Image Processing

Background:

• Binarization is crucial for object recognition in image analysis.
• Iterative segmentation methods are effective but struggle with stopping criteria.
• Accurate segmentation directly impacts the performance of image analysis systems.

Purpose of the Study:

• To introduce a new algorithm for medical image binarization.
• To utilize entropy as a stopping criterion in image segmentation.
• To improve the quality of binarization in medical imaging.

Main Methods:

• Recursive application of mean shift filtering for image segmentation.
• Using image entropy as the stopping criterion for the segmentation process.
• Performing binarization after obtaining the segmented image.

Main Results:

• The proposed algorithm demonstrates good binarization performance on medical images.
• Experimental results show high-quality binarization compared to existing methods.
• The new method was compared against a previous algorithm and Otsu's method.

Conclusions:

• Entropy-based stopping criterion effectively guides iterative segmentation for binarization.
• The novel algorithm offers an improved approach to medical image binarization.
• The method provides a reliable solution for enhancing object recognition in medical imaging.