A novel nonparametric technique for segmenting multimode hyperspectral images obtained from non-melanoma skin cancer lesions

Abstract

Keratinocyte Carcinoma, more traditionally known as Non-melanoma skin cancer (NMSC), is the most common cancer in humans. Incidence continues to increase despite increased public awareness of the harmful effects of solar radiation. In this paper, a non-parametric technique based on image registration will be applied to the multimode hyperspectral imaging system to segment Basal Cell Carcinoma (BCC) and Squamous cell carcinoma lesions (SCC). The aim is to enhance Mohs surgery by determining the actual borderlines of the desired area in the patient's images, leading to increased efficiency and efficacy of the Mohs surgery. The proposed algorithm was applied to four sets of different Multimode hyperspectral Images with Non-Melanoma Skin. The experimental findings showed that the proposed algorithm is effective in Non-Melanoma skin detection. This could lead to improved image-guided excision of cancerous lesions with potential applications in robotic interventions.