DEEP HASHING FOR SECURE MULTIMODAL BIOMETRICS

Abstract

A multi-biometric method was employed to improve the accuracy of the 
authentication process while simultaneously lowering mistake rates. Many systems, such as 
access control, PC login, e-commerce, and so on, need person identity. The biometric system 
is most likely utilized for security. The two frameworks of biometric systems are unimodal 
biometric and multimodal biometric. 
In a unimodal system, a single biometric feature is employed, while a multimodal system uses 
many biometric traits. 
In comparison to a single-modal biometric framework, a multimodal biometric framework is 
more exact. This proposed study will address the many types of biometric systems, such as 
unimodal and multimodal systems. Discuss the comparability of several prior modalities in 
biometric systems and their comparative analysis. A multi-modal system is used to compare 
receptive techniques. The need for biometric systems is increasing on a daily basis. The 
disadvantages of the unimodal system are also shown, which explains why the need for 
multimodal transportation will expand. During this analytical task, we will mostly analyze 
earlier work that is unimodal and multimodal. Two features, such as fingerprints and iris scans, 
are merged in the proposed multi biometric system. The suggested system is evaluated using a 
standard database. Various characteristics are extracted from each trait using various feature 
extraction methods. 
The matching score of these extracted characteristics is determined independently. The 
weighted fusion approach is used to integrate these separate scores. According to the 
observations, 96% accuracy is attained, overcoming the constraints of the current method. 

,