A Finger Vein Image-based Person Recognition System using BCH Codes to Access Medical Records

In recent days, Biometric features like iris, face, fingerprint, voice pattern, and gait are wide-ly used in private and public sectors to secure data for identifying a person. Finger veins of an individual are hidden biometric attributes that cannot be accessed easily and it is even different in identical twins. In this paper, personal identification is done using a finger vein to access medical reports. Initially, a finger vein image is extracted from personnel and preprocessed. The preprocessing is done by resizing the image and applying the Sobel filter as an edge detection operator. Then, the Maximum curvature method and repeated line tracking methods are used to identify the finger vein patterns. Then, the Gabor filter is utilized to ex-tract features and analyze the texture. Then, the feature-extracted image is compressed using the principal component analysis method. The compressed image is encrypted and decrypted using a fuzzy commitment scheme along with error-correcting codes to improve accuracy. The proposed system performance is evalu-ated by accuracy, sensitivity, and specificity and its values are 96.15%, 96.7%, and 95.2%. The proposed work achieved the accuracy on par with existing state of art methods.