Abstract
A facial recognition system for designated members which including based on detection, facial feature points location and facial recognition were developed. Robust feature location algorithms were adapted to discreminate facial features location precisely, the feature points were the transformed into feature vectors and then compared with database. In other words, to correlate each feature vector, e.g. distance, area… etc, comparison of the data of unknown facial image with database using Euclidean distance algorithm was adopted, and recognition results can be obtained. To achieve a higher recognition rate, modified Euclidean distance was proposed and implemented. Furthermore, Linear Discriminant Analysis Method was adopted to classify the data more accurately. Finally, we combined the modified Euclidean distance and Linear Discriminant Analysis method in the recognition process. A better than 90% recognition rate for specific group of people can be achieved through experiments.