Level set method for segmentation of infrared breast thermograms.

Area of Science:

• Medical Imaging
• Biomedical Engineering
• Oncology

Background:

• Breast thermography is a non-invasive physiological test measuring breast surface temperature distribution via infrared radiation.
• Increased temperature in precancerous or cancerous tissues, due to angiogenesis and heightened metabolic activity, makes thermography a potential early detection tool.
• Thermography can identify early abnormal changes before they are detectable by mammography.

Purpose of the Study:

• To evaluate and compare the efficacy of three image segmentation methods (k-means, fuzzy c-means, and level set) for analyzing breast thermography images.
• To assess the ability of these methods to accurately extract suspected regions of interest indicative of abnormalities.
• To determine the most accurate segmentation technique for identifying tumor shapes in thermal breast images.

Main Methods:

• Breast thermography images were analyzed using k-means, fuzzy c-means, and level set segmentation techniques.
• The methods were applied to various cases, including fibrocystic and inflammatory breast cancer.
• The extracted 'hottest' regions were compared against original thermal images to evaluate accuracy.

Main Results:

• All three segmentation methods were capable of extracting hot regions from thermal breast images.
• The level set method demonstrated superior accuracy in segmenting these regions compared to k-means and fuzzy c-means.
• The level set approach showed significant potential in extracting the precise shape of tumors.

Conclusions:

• Image segmentation techniques are crucial for extracting relevant information from breast thermography.
• The level set method is a more accurate and promising approach for the detection and characterization of abnormalities in breast thermography.
• Accurate segmentation of thermal patterns can enhance early breast cancer detection capabilities.