FACE RECOGNITION AND DETECTION ON IMAGE ENCHANCEMENT TECHNIQUES

 Abstract

Our aim is to detect and recognize faces in Digital Images. We have implemented our idea to train the computer to “see” or “know” how a typical face will be, a computer vision problem. Apart from using Neural Networks, we have used “Principal Component Analysis”(PCA)/KL transform method combined with several distance measures in n-dimensional space to implement a face detection system. A test image is then scaled down into several smaller images and a sliding window is dragged across each image.
 The various distance measures are then used to classify the area inside the window as a face or not. Distance measures namely Euclidean distance and Mahalanobis distance have been used. These distance measures have been used for “Maximum Face Likelihood Detection”.For better results ML detection has been combined with “Symmetricity detector”.We have shown the “Performance results of our implementation” which we have done. Face detection has been considered to be the first step towards our intended project on comprehensive “Emotion Recognition” from facial features. We have also implemented new Image Enhancement techniques by an Image Algorithm”. We have simulated the results using “MATLAB”.

Click here to download the full paper