Human skin colour detection is a necessary element in computer vision as well as in image processing applications. It is a separation procedure of skin pixels from non-skin pixels in image. Whereas, detecting human skin colour is a very difficult function due to two reasons, mainly, changeable illumination conditions and diverse races of people. Some previous researchers in this field tried to resolve those problems by using thresholds that relied on specific values of skin tones. Although, it is a speedy and an easy implementation, it does not provide sufficient information for recognizing all skin tones of humans. This paper proposes Bayesian Rough Decision Tree (BRDT) classifier to improve the accuracy of human skin detection. Three experiments have been conducted using (RGB) dataset collected from University of California, Irvine (UCI) machine learning repository, RGB (Red, Green, Blue), HSV (Hue, Saturation, Value) and YCbCr (Luminance, Chrominance). The experimental result shows that the proposed system can achieve preferable accuracy in skin detection 98%, 97% and 97% using RGB dataset, HSV dataset and YCbCr dataset respectively.
CITATION STYLE
Abbas, A. R., & Farooq, A. O. (2018). Human skin colour detection using Bayesian Rough Decision Tree. In Communications in Computer and Information Science (Vol. 938, pp. 240–254). Springer Verlag. https://doi.org/10.1007/978-3-030-01653-1_15
Mendeley helps you to discover research relevant for your work.