Bag-of-Features based classification of Dermoscopic Images

Dermoscopy image analysis is one of the open challenge and important issue related to biomedical image analysis. Determination of different diseases of skin is necessary in medical practices. Automated techniques are proved to be useful in detection and classification process of different skin diseases. There are several automated computer guided tools and frameworks have been proposed for early detection and analysis of skin diseases. In this work a comprehensive study on dermoscopic image analysis using bag-of-features has been presented. Bag-of-features is one of the well-known classification method. It uses local description of images. In case of dermoscopy image analysis, color and texture plays a vital role. Color and texture descriptors have been also discussed. Future directions in dermoscopy image analysis based on bag-of-features model are also presented.