Consideration for segmentation based on radiometric data processing, towards the research of quantitative medical thermography

Medical thermography is considered an auxiliary technique in preventive medicine, treatment monitoring, and surgery. However, the true value of thermography lies in quantitative studies so that the information presented to the human eye can be maximized. To this end, it is essential to preserve the radiometric data that may provide information to quantitative studies, e.g., temperature difference analysis, segmentation, contrast adjustments, artifact removal, etc. However, certain thermographic equipment may not be radiometric so only a mapped thermal image known as a false color image can be obtained. Therefore, the radiometric aspect must be considered to perform quantitative studies. This paper presents the comparison between processing radiometric data matrices and false color images. The methods employed were segmentation of the region of interest (RoI) and contrast adjustment once segmented. The segmentation method is based on thresholding, in which an optimal threshold was found to be approximately 0.8 for radiometric arrays and 0.6 for grayscale images that were not transformed from an RGB image. In conclusion, using radiometric data arrays has advantages over using RGB images as input information. Moreover, medical thermography should not only consider a qualitative aspect when it comes to diagnostics, but the validation of thermography lies in quantitative studies based on radiometric input.