Dermoscopic Image Enhancement and Hair Artifact Removal using Gabor Wavelet Uzma Jamil Member IEEE, Ph.D Scholar, Department of Computer Engineering, Bahria University, Islamabad, Pakistan. Senior Lecturer, GC University, Faisalabad, uzma_gcuf@yahoo.com Dr. Shehzad Khalid Professor, Head, Department of Computer Engineering, Bahria University, Islamabad, Pakistan. shehzad_khalid@hotmail.com Dr. M.Usman Akram Assistant Professor, National University of Sciences & Technology, Islamabad, Pakistan. usmakram@gmail.com Abstract— Skin cancer is one of the most prevalent types of cancer in our world. Diagnosis of skin cancer needs specialized equipment, doctors and continuous monitoring. Patients living in remote areas normally cannot access such facilities. To overcome these barriers of access, Computer Aided Diagnostics, an emerging field in computer science, often called telemedicine, is being considered a promising approach. Image processing for Computer Aided Diagnostics has three key steps, i.e. Segmentation, Feature Extraction and Classification. In this research, preprocessing and hair artifact removal experiment was performed on dermatoscope images by using Morphological and Gabor wavelet-based techniques. It has been found that, in some cases, wavelet transformations provide better results as compared to other techniques like gel, water bubbles and dark hair around the surface affected by cancer, i.e. these artifacts are removed with less effort. Experiments also showed that images with Blue channel from RGB are better as compared to other grayscale conversion techniques. Index Terms— skin cancer, dermoscopy, image enhancement, hair Inpainting, hair segmentation. I. INTRODUCTION Computer added diagnostic (CAD) system has three important steps i.e. image segmentation, feature extraction and classification [1]. In case of skin lesion the homogeneity will be the properties of effected part with each other or homogeneity between the normal part of skin. So the purpose of image segmentation of skin lesion is to segment the effect part of skin as close as possible from the normal skin. Better segmentation leads better features extraction, classification and diagnosis [12]. Thus diagnostic system's final results are highly dependent on the quality of images that excludes maximum possible noise types. In order to perform skin lesion segmentation effectively some predefined post and pre- processing steps are required and these steps are discusses in detail by Abbas [2]. As dermoscopic image contain some artifact like hairs, gel, skin lines, dark spots and water deposits which can affect the segmentation results. In order to handle these artifact some per-processing steps of segmentation are applied and then perform the segmentation. After the segmentation the image may also contain some non-reliant information like small island or irrelevant object and those also needed to be removed by post processing steps. And Abbas discussed these steps in detail. Below Fig.1 is showing different sample images of melanoma skin cancer. Fig. 1. Sample dermoscopic images containing malignant melanoma and highlighting the variations in the melanoma lesions (a) Smooth transition between lesion and the image; (b) Specular reection; (c) Presence of hairs; (d) Gel and bubble presence There are number of skin cancer types however melanoma is the most aggressive, complex and deadly one. Survival rate of melanoma is highly depended on its early detection and treatment . The manual detection of melanoma in its early stage depends on the availability of expert doctor with proper medical facilities and equipments it is thus neither cost effective nor time efficient. The situation get more worst for skin cancer in underdeveloped or remote areas due to lack of access to proper medical facilities. As an alternative, Computer Added Diagnostic (CAD) has started offering better solution in