A Rotation-Invariant Fingerprint Identifications System Using Neural Network

File Size:
1.86 MB
Volume 3, Issue 8 (August, 2017)
Publication No:
Pushpinder Kaur, Er. Khyati Marwah
6 x

Fingerprint identification and confirmation systems provide automatic acknowledgment of an individual grounded on certain unique topographies or characteristics possessed by that individual. It takes into consideration the natural features of every single individual for various applications in today’s fast growing technical world. Fingerprint systems also help in overcoming the existing problems and limitations in authentication and identification fields and prove efficient and accurate in security related issues. Various biometric aspects are used for developing such systems like voice, ear, heartbeat, signature, face, fingerprints etc. Iris recognition is measured as one of the most accurate biometric methods available owing to the unique epigenetic patterns of the fingers. A system is developed that can recognize human fingerprint patterns and an analysis of the results can be done. Fingerprint has unique genetic combination for every individual that remains same for a long period throughout one’s lifetime. The genetic combination of a fingerprint does not match with any person on this earth. This special feature of fingerprint is used in the biometric system presented in the research work. A noble mechanism has been used for implementation of the system. Fingerprint recognition system is implemented by following defined steps like image acquisition, edge detection, segmentation, feature extraction, optimization and matching. Images are taken from the database that contains gray scale images for pre- processing. In the pre-processing step, the fingerprint image is segmented using canny edge detection method and histogram equalization. Later, an efficient extraction algorithm based on optimized Gabor filters is done. On the extracted feature vector, minutiae’s are extracted for fingerprint recognition. The feature vector is fed to the neural network as a training vector. At the end, test sample is matched with the sample image of fingerprint using neural network classification technique. Experimental results are performed using performance parameters named as False Acceptance Rate, False Rejection Rate and Accuracy of the system.

Fingerprints, iris, False Acceptance, False Rejections, minutiae, Biometric, DNA.