Main Article Content
Abstract
The great usage of mobiles, Internet and social media give the opportunity for the existence of trillions of images on the World Wide Web beyond our imagination. This created a substantial demand for algorithms to perform content-based image retrieval of images. Now researchers, are concentrating on the identification images likefully dressed, partly dressed or no dressedwhich is also become vital area of research called natural content-based retrieval problem. The skin detection is the key research in order to study the image portions for the efficient categorization. This paper uses color features based image attributes using simple histogram learning techniques. We used the Fashionista benchmark dataset v0.2 and tested with 25 random images. The results are surprisingly satisfied and achieving a considerable success rate of 85% true positives and 15% true negatives. We believe that this work will be improved further to achieve the goal. The entire work is done using Matlab 6.0.