HSV Color Model and Logical Filter for Human Skin Detection

Abstract

Since color is unaffected by scaling, rotation, or partial occlusion, it is an effective function for object detection. The identification of skin color is an essential stage in an eclectic range of computer vision applications. The rapidly expanding field of detecting human skin is focused on the idea that colored images can be used to collect details concerning individuals, mode, purpose, and image contents, and computers can then react appropriately. Skin color of humans can differ significantly in manifestation due to variety of aspects such as lighting, ethnicity, ageing, imaging conditions, as well as complex context, making human skin detection in complex images a difficult issue. Many techniques, on the other hand, have been created to deal with the issue of skin identification in color photographs. The skin detection algorithm proposed in this research paper is based on thresholding HSV color model bands with upper and lower bounds to create a correct mask, which is then logically ANDed with the original image to restore the skin tone pixels. For images taken in the natural world and under various illuminations and light conditions, the experimental results showed 99.135 percent precision and 99.587 percent accuracy.