Computer aided diagnosis of skin lesions

Date

2012

Authors

Sanchez, Isaac Armand

Journal Title

Journal ISSN

Volume Title

Publisher

Abstract

Melanoma is the most deadly form of skin cancer. The World Health Organization estimates that more than 65,000 people a year die worldwide from too much sun, mostly from malignant skin cancer. Early diagnosis of melanoma is essential for the fight against skin cancer and if detected early enough, physicians and specialists can treat the skin cancer while it is still localized in the originating skin lesion. While dermatological methods are objective, the scoring itself is by nature subjective and based on the physician's opinion. Hence, accurate detection, and early prediction of cancer is imperative to medical research today. The visual recognition by clinical inspection of the lesions performed by dermatologists is around 75% accuracy. This thesis develops computer aided methods of detection and diagnosis of skin lesions. It discusses extracting lesion images from a larger image of a skin surface and evaluates the ability of feature sets to classify the extracted skin lesions. In order to accomplish this task, this thesis reviews methods of dermoscopy, segmentation, feature extraction, feature selection and finally classification methods. This thesis includes a review of existing features and an introduction of new features based on 2-dimensional color histograms and the shape-adaptive discrete cosine transform. Combined with an automated segmentation method using a fusion of thresholding methods, the resulting system is designed to be used by dermatologists as a complete integrated dermatological analysis tool to improve the rate of correct diagnosis well above 90%. Simulations are implemented to show extraction of skin lesions and how their features are measured as well as classification experiments. The outcomes showed that the CAD models discussed in this paper have an improved classification performance and are objective diagnostic tools that can be used in medical practice.

Description

This item is available only to currently enrolled UTSA students, faculty or staff. To download, navigate to Log In in the top right-hand corner of this screen, then select Log in with my UTSA ID.

Keywords

Classification, Feature Extraction, Image Processing, Skin Cancer

Citation

Department

Electrical and Computer Engineering