Pioneering Skin Health: Harnessing Thermal Images for Cutting-Edge Computer-Aided Diagnosis

Skin diseases are one of the most common health problems across the world. The detection of skin diseases has a huge impact on treatment. The non-invasive method, currently being followed is, through visual examination of dermoscopic images of the lesion by a physician with the knowledge and technical experience in the relevant field. In some cases, human errors in diagnosing disease may lead to false treatment, which may be fatal. Automated detection through Dermoscopy has been in the research and is not being used in practice due to a few limitations such as inaccurate output for dark-skinned patients, and the presence of air bubbles and hairs. on the skin, detection of disease with the help of superficial layer image may have false accept error. Thermal images, obtained with the help of infrared or thermal cameras are believed to eliminate these errors. The infrared radiation camera can detect the presence of any kind of lesion around a specific region of skin and can efficiently reproduce the results. Temperature distribution around the skin is measurable and the abnormality can be easily quantified using a thermal camera. In the proposed work, MATLAB was used to evaluate 100 data samples, which had been acquired. Out of these, 34 were determined to have severe lesions, 22 to exhibit moderate lesions, 20 to possess normal lesions, and an additional 30 had been selected as controls for the normal group. The obtained results of the work demonstrate the efficiency of processing thermal images for the binary classification of skin conditions.