Abstract
This project is about human face recognition in image files and human face detection in video streaming. Face recognition involves matching a given image with the database of images and identifying the image that it resembles the most. Face detection is a computer technology that determines the locations and sizes of human faces in digital images. In this project, face recognition is done using two methods: (a) Eigen faces and (b) applying Principal Component Analysis (PCA) on wavelet sub-band image. In the first face recognition method, Eigen faces technique is used. Development is done in Java. The second method for face recognition here is to apply Principal Component Analysis (PCA) on wavelet sub-band image. Instead of using the entire image for creating the test database and sample test image, the images are decomposed into frequency sub-bands. This provides greater computational speed when the test database is huge because low resolution decomposed images are used instead of high resolution images. This is also done in Java using standard Java Graphical User Interface packages for Wavelet Transform. For face detection, Haar-like feature based method is used here. Haar-like features are digital image features used in object recognition that encode information about the object to be detected. Java Media Framework (JMF) library is used to show the video captured from the web camera in a Java application. Java imaging libraries are used to develop the algorithm for the face detection. The aim of this project is to successfully demonstrate the human face recognition using Eigen faces and Wavelet sub-band methods and also to detect human face in a video streaming.