ABSTRACT

There is no tool to block pornographic images included in the PDF documents. So, this paper proposes a system to extract images from PDF documents, and detect the skin color regions in the extracted images as a step to filter undesirable images from PDF documents. We studied two famous skin color detection techniques. The first technique is based on the log opponent color space and the second technique is based on the HSV color space. We modified the two techniques to overcome the failure of detecting complex image’s background by combining each of the two techniques with the YIQ skin detection technique to obtain accurate skin detection results. Finally, the results of comparing the two skin detection techniques in terms of the number of pixels detected are presented.