A Novel Method for Finger-Vein Recognition Based on MFAGNET Approach

Biometrics based on veins are gaining prominence due to their higher level of accuracy and security. A more advanced biometric solution to the problems with fingerprint systems is the use of finger veins. A promising biometric pattern for personal identification, vein-based authentication offers both security and convenience. To acquire the vein patterns, one must use a live subject. Thus, by means of natural and convincing evidence, it is demonstrated that the patient whose veins were successfully extracted is alive. Be careful to maintain the importance order while completing preprocessing, feature selection, and training the model. The proposed approach employ Harris Corner in preprocessing. In images, corners are areas where there are noticeable changes in brightness in more than one direction. Assuming there is enough data, the objective of LLE in feature selection is to locate a low-dimensional embedding. Unified MFGANET models need to be trained after feature selection. Current state-of-the-art methods, CNN and LSTM, are outdone by the suggested methodology. Following the technique's use, accuracy improved by 98.16%.