Main Article Content
Melanoma is a type of skin cancer. Dermoscopy is a non-invasive technique to examine skin lesion. It helps in detection of melanoma. Experienced dermatologists are required to diagnose melanoma from dermoscopic images. Death toll of melanoma patients is very high due to difficulties faced in diagnosis. Computer-aided design is a way to diagnose it at early stage and hence proper treatment will lead to decreased death rate. Digital dermoscopic images are processed to diagnose and classify skin cancer. Noise present in image is a hurdle in the detection of lesion. Gabor wavelets are used to highlight artifacts. Hair, reflections, shadows, skin lines and air bubbles are artifacts that are considered as noise. Noise degrades image quality. The corrupted image can lead to wrong detection of cancer region. Image visibility can be enhanced by improving color contrast and noise removal. Image processing techniques are applied to images to remove artifacts. This paper presents comparative analysis of ten filters and six point processing techniques. De-noised image is further processed for feature extraction, segmentation and classification.
Copyright (c) 2018 Technical Journal
The author transfers all copyright ownership of the manuscript entitled (title of article) to the Technical Journal in the event the work is published.