Thermal image segmentation based on density slicing of color histogram of images

Abstract

Thermal imaging has been recently used as a new approach for human biometrics. The formation of thermal image is purely based on the heat distribution of an object. It brings some difficulties to image segmentation due to its over centralized intensity distribution and low intensity contrast. Our approach uses color slicing segmentation method to characterize the texture information using various color models.

The experimental results of present study shows the color slicing technique can make good effect on the segmentation of tumor structure. We have found that segmenting these images can be made easier by using RGB, HSV, YCbCr color models by Color slicing method. We have used many Statistical operations (such as contrast, variation …) to evaluate our segmentation results.