Application of fishface algorithm to face recognition system

摘要

Face recognition system for the student credit card application requirements, selection Fisherface methods for image processing. Process: a system for training sample S, by S to find a recognition of complexity and can reduce noise reduction transformation, to be marked with the name of this face image is stored in the database after conversion. When you want to identify a face image, the image transformation processing same, with the results obtained in the database samples of each individual face comparison, find the Euclidean distance or Mahalanobis distance, and the minimum distance corresponds to that person The name is output. Algorithms and ported to ARM embedded system, to achieve a dynamic collection of students and face recognition, face samples of 30 individual tests, the recognition rate of 87.502%.