Finger knuckle biometric feature selection based on the FIS_DE optimization algorithm

In recent years, the hand-based biometric system has received significant attention in identifying a person. In the hand-based biometric, the finger knuckle print plays a vital role in recognizing a person. The main purpose of this research is to propose an effective feature optimization technique for identifying the best feature vectors for finger knuckle print-based authentication. This work presents a novel feature selection algorithm, fitness index selection with differential evolution (FIS_DE), based on K-nearest neighbor (KNN). Initially, the feature extraction is performed using the conventional methods like principal component analysis, linear discriminant analysis, and independent component analysis. Then, evolutionary algorithm is used for feature selection with best vectors. The population-based metaheuristic algorithm DE proposed is used to optimize the KNN classifier. FIS_DE_KNN is compared with Euclidean and neural network classifiers to show the improved efficiency of the proposed work. In this research, experimental results of the proposed FIS_DE-KNN improve the classification accuracy with an optimized number of features.

Automated summary

This research proposes an effective feature optimization technique using the K-nearest neighbor algorithm and differential evolution for finger knuckle print-based authentication. Experimental results show improved classification accuracy with this method.