YOLOv8’s head-layer Performance Comparison for Skin Cancer Detection

Abstract

Skin cancer is a type of cancer that can lead to death. The mortality rate from this disease is high. Detecting the disease at an early stage is essential to prevent the worst impact. However, detection by a dermatologist is time-consuming and costly. Computer Aided Detection (CAD) systems are that aid dermatologists in the understanding of medical images. Deep learning strategies are repeatedly employed in CAD systems. Yolo is one of the well-known deep learning models used to solve detection cases for small, medium, and large objects. In this study, we evaluate the performance of Yolov8 for skin cancer detection considering three head-layers on the HAM10000 dataset. Experimental results show that the head layers with large object paths produce the best mAP and significant speed compared to medium and small. With these results, it can be considered as a reference in future research that using the big object detection path to detect skin cancer at an early stage is recommended.