Object Recognition Using Eigen Values

Object recognition is a computer application for automatically identifying an image from a set of images in the database. Many recognition systems recognize the images based on some characteristic features of the image. However, this paper helps recognize the images based on some mathematical computations which include the Eigen values and the Eigen vectors. The Eigen face approach is one of the simplest and most efficient methods in recent times for developing a system for Object Recognition. Object recognition is an emerging field with applications varying from video surveillance, forensic use to commercial applications such as in virtual reality, smart cards and information security. Eigen faces are eigenvectors of covariance matrix, representing the given image space. The object images are projected onto a face space (feature space) which best defines the variation of the known images. The face space is defined by the ‘Eigen faces’ which are the eigenvectors of the set of images. The new image which is projected into this face space is then compared with the available projections of the training set to identify the image. The image with a minimum distance (less than the threshold distance) from the input image in the projection area is most suited to be the result. The framework provides the ability to learn to recognize objects in an unsupervised manner. New images which are fed into the system can be identified with a high rate of success.