The Random Forest Classifier Applied In Biometric Recognition

Abstract

Biometric recognition technology offers a high level of security and protection, mak- ing identity verification easy and fast, while reducing human errors and enhancing verification accuracy. In this study, we proposed a biometric recognition system based on minor finger knuckle (MFK), because it has high accuracy, reliability, and resistance to tampering. This technology is beneficial in various fields such as security, access con- trol to buildings and devices, secure payment applications, and identity recognition in mobile devices. The Discrete Cosine Transform (DCT) technique was used for feature extraction, and the Random Forest (RFT) technique was used for features classifica- tion. A proposed multi-modal system was developed, and its performance evaluated by using matching score level fusion. The evaluation was conducted on a database con- taining fingerprint images of 500 individuals. From different experiments, obtained results were presented by using unimodal and multimodal recognition systems. Based on the results, excellent performance was achieved in multimodal biometrics recognition compared to unimodal biometrics.