A SYSTEM AND METHOD FOR INTEGRATION OF BIOMETRIC RECOGNITION OF VARIABLE COMPUTING COST AND COMPLEXITIES

The present invention provides a framework for integration of biometric recognition methodologies of variable computation cost. The system of the present invention comprises a Detection Module (102) which captures images through a camera from a user and detects biometric data through a browser at the client platform; an Image Processing Module (104a and 104b) which enhances the quality of ROI images through relatively fast and relatively slow processing which will enhance the quality of biometric image; a Feature Extraction Module (106a and 106b) which extracts facial features by means of multiple methods which range from relatively computation-inexpensive to relatively computation- expensive from enhanced ROI images; a Feature Matching Module (108a and 108b) which calculated distances between the multiple feature vectors previously extracted and the multiple feature vector which has been stored in database; and an Authentication Module (110) which combines multiple distance scores obtained from Feature Matching process in order to undertakes decision of acceptance or rejection of user based on submission of applicable video-stream. In the present invention, biometric data is captured by means of an apparatus (i.e. camera) attached to the client platform. Upon capturing biometric data by utilizing camera, the processes of biometric detection, signal processing and feature extraction are executed. Concurrent execution of expensive and inexpensive computation would respectively result in relatively slow and relatively fast computation of the biometric feature vectors associated with the user of interest. The present invention allows integration of these relatively fast and slow assessments of biometric authenticity, and also for multiple progressive assessments based on the relative speed of these computations. (The most illustrative drawing is FIG. 1.0.)