A. Eye Detection Using Varients of Hough Transform B. Off-Line Signature Verification

PART (A): EYE DETECTION USING VARIANTS OF HOUGH TRANSFORM: Broadly eye detection is the process of tracking the location of human eye in a face image. Previous approaches use complex techniques like neural network, Radial Basis Function networks, Multi-Layer Perceptrons etc. In the developed project human eye is modeled as a circle (iris; the black circular region of eye) enclosed inside an ellipse (eye-lashes). Due to the sudden intensity variations in the iris with respect the inner region of eye-lashes the probability of false acceptance is very less. Since the image taken is a face image the probability of false acceptance further reduces. Hough transform is used for circle (iris) and ellipse (eye-lash) detection. Hough transform was the obvious choice because of its resistance towards the holes in the boundary and noise present in the image. Image smoothing is done to reduce the presence of noise in the image further it makes the image better for further processing like edge detection (Prewitt method). Compared to the aforementioned models the proposed model is simple and efficient. The proposed model can further be improved by including various features like orientation angle of eye-lashes (which is assumed constant in the proposed model), and by making the parameters adaptive. PART (B): OFF-LINE SIGNATURE VERIFICATION: Hand-written signature is widely used for authentication and identification of individual. It has been the target for fraudulence ever since. A novel off-line signature verification algorithm has been developed and tested successfully. Since the hand-written signature can be random, because of presence of various curves and features, techniques like character recognition cannot be applied for signature verification. The proposed algorithm incorporates a soft-computing technique “CLUSTERING” for extraction of feature points from the image of the signature. These feature points or centers are updated using the clustering update equations for required number of times, then these acts as extracted feature points of the signature image. To avoid interpersonal variation 6 to 8 signature images of the same person are taken and feature points are trained. These trained feature points are compared with the test signature images and based on a specific threshold, the signature is declared original or forgery. This approach works well if there is a high variation in the original signature, but for signatures with low variation, it produces incorrect results