Content Based Video Retrieval for Obscene Adult Content Detection

With the advancement in networking, content produced and distributed over the Internet is exponentially increasing. This imposes the threat of distribution of obscene content freely and largely, urging mechanisms to control access by minor aged users. Manual retrieval and indexing of material is impossible for large video repositories. This paper proposes a method to detect videos with obscene adult content using content based video retrieval techniques. We propose an algorithm to summarize the video by extracting keyframes that mark video shot boundaries and apply BoVW algorithm to classify keyframes indicating the presence of obscenity. Despite the ignorance of high-level features in temporal domain, a higher recognition rate of 85% with spatial information alone is proved. Further, we show the irrelevance of color information to detect nudity in videos when using BoVW.