Normal moles are mostly small brown in color that are spots or growths on the skin that is by birth or emerge in the first few decades of life in almost everyone. Skin cancer most often appears as moles. Seldom people are aware about the skin cancer. If detected at early stages then it can be cured. We propose an improved design of Gabon filter which is combined with Vedic algorithm (Urdhva Triyagbhyam) to give faster convolution result. The exertion has proved the efficiency of Urdhva Triyagbhyam which is the Vedic method of multiplication that enables parallel generation of intermediary products that eliminates unwanted multiplication steps with zeros and scaled to higher bit. In this paper we employ the filter design appropriate for detecting the early stages of skin cancer using textural properties of skin. © 2015 IEEE.