Pornography and other such obscene material are now easily available on the internet, social media and other video and document sharing sites. Tons of videos are uploaded and shared daily on the Internets that have made it impossible to tag and filter obscene matters manually. Only feasible solution is automatic system that can detect pornography other similar materials in multimedia formats. Unfortunately most of such solutions are complicated and computationally expansive. Most of these solutions work with high-level features either special or in time space. On the other hand, low-level features are easy to detect and are fast in the computations, but contain less contextual information. In this article, a set of low-level features are analyzed to find the possibility of their use in pornography detection. Experimental results show that most of these low-level features are not suitable in this domain. However if different ratio of the same features (mix of these features) are used intelligently, they can be successfully used for the high-level content recognition as nudity and pornography detection.
Weitere Kapitel dieses Buchs durch Wischen aufrufen
Bitte loggen Sie sich ein, um Zugang zu diesem Inhalt zu erhalten
Sie möchten Zugang zu diesem Inhalt erhalten? Dann informieren Sie sich jetzt über unsere Produkte:
- RGB and Hue Color in Pornography Detection
Neuer Inhalt/© ITandMEDIA, Product Lifecycle Management/© Eisenhans | vege | Fotolia