Increasing the reliability of skin detectors

Skin detection is a well-known image processing technique that has been used in many applications such as video surveillance, naked image filters, and face detection. This paper proposes a reliable skin detection method that integrates both color and texture features. To increase the reliability of the skin detection process, neighborhood pixel information is incorporated into the proposed method. Texture features were estimated using statistical measures as range, standard deviation, and entropy. Back propagation artificial neural network is then used to learn features and classify any given input. In this work, three skin detectors based on pre-defined rules of skin color tones, texture features only, and a combination of both color and texture features have been constructed. Furthermore, the paper analyzes and compares the obtained results from the proposed skin detector with pre-defined color tones and texture feature-based skin detectors to show the level of robustness of the proposed skin detector. It found that the proposed skin detection method achieved a true positive rate of approximately 95.6176% and a false positive rate of approximately 0.8795%. Experimental results showed that our approach is more efficient compared with other state-of-the-art color-based or texture-based skin detector approaches