Evaluation of Performance of the Face Detection Using Skin Colour Model and Face Recognition Using ANN

This paper presented a new and robust algorithm for face detection and recognition using skin colour model with Feed Forward neural Network. Skin colour model is used to detect face of region from given input testing image. After detection, Features are extracted from detected face regions using Zernike moment and correlation. Also Extract features from training database. Next step is applied to fuzzy set for pre-processing of extracted features of training image. Fuzzy set has been short listed more match able images (greater than min threshold value of matching feature).The system has been tested on face pix database in which 6600 images of 90 persons of 74 different images of angle variation. On evaluate performance of new algorithm on basis of factor such as accuracy, time, MSE etc. New System was found to have overall accuracy 94.64% at distance 60 to 80 cm of images. Keywords—Face Detection, Face Recognition, Artificial Neutral Network (ANN), Skin colour model, Zernike moment