Skin Cancer Segmentation Using Neural Network: Image Processing Methods and Kohonen Neural Network

Skin cancer is the most common form of cancer in United States. If detected at an early stage simple and economic treatments can cure skin cancers. Accurate skin lesion segmentation is critical in automated skin cancer early detection and diagnosis. This book proposes methods like explicit pixel based, median filter, difference filter, conservative smoothing and Kohonen neural network for automated segmentation which can replace traditional biopsy and dermoscopes. Further it can give direction to dermatologists in detecting ABCD's of skin cancer. Also may help research scholars, practitioners and students to understand more about image processing methods and role of neural network in segmentation.