Tattoo skin detection and segmentation using image negative method

Tattoos, a soft biometric trait, are gradually being used for suspect and victim identifications in forensics and law enforcement. Particularly, tattoos are raising obvious evident attention because of their visual and demographic traits as well as their increasing prevalence. However, tattoos on human skin are complicated and large invariance in both structure and skin surface. In order to improve tattoo image retrieval and matching, this paper proposes an approach of tattoo skin detection and segmentation using the image negative method in the pre-processing part. The process is composed of three steps. The first one is the skin detection where we use a variety of skin patches to do the task of human skin colour segmentation using the HSV model, especially, Asian skin colour. Then, in the second step, the image negative method is used for detecting the clear graphic image of the tattoo segment. Finally, we extract the tattoo segment from the skin area of the negative image and, as a result, the tattoo negative image is obtained and can be used for retrieval. Our experimentation has been carried out based on the dataset of tattoo images, gathered from Thai Criminal Records Division - Royal Thai Police, Kingdom of Thailand. Based on the concept of CBIR (Content-Based Image Retrieval), SIFT (Scale Invariance Feature Transform) has been employed in the process of image matching and retrieval. The result has illustrated that the tattoo skin detection and segmentation are efficient and effective for tattoo image retrieval, and, also, reduce the possibility of illogical matches.