110 likes | 301 Views
Face Detection in Color Images. Method Used . Face detection using skin color based segmentation. Data Set . A Data set of more than 100 singular face images with faces directly facing the camera to get a better skin color estimation. Source :- Google and AltaVista search engines.
E N D
Method Used • Face detection using skin color based segmentation
Data Set • A Data set of more than 100 singular face images with faces directly facing the camera to get a better skin color estimation. • Source :- Google and AltaVista search engines
Steps Followed • Use of HSV color space for color segmentation instead of RGB. • Training of the algorithm using the given dataset in HSV color space. • Using Images with multiple face images to generate the final results.
Results obtained • For Skin Color classification the algorithm gave the following results: 17/255 <= H <= 240/255
Original Image Segmented image
Original Image Segmented Image
Original Image Segmented Image
Future Work • Getting Better approximation of skin color using factors like saturation and Value in HSV and also using other models such as YCbCr for further refinement. • Extraction of actual faces from the segmented images using symmetric properties of face and elliptical objects
References and Platforms • Matlab Image processing tool box. • Detecting Human Faces in Color Images by Ming Yang and Narendra Ahuja. • A Bayesian approach to skin color classification in YCbCr Space by Douglas Chai and Abdesselam Bouzerdoum.