An Economical, Interpretable, and Scalable Biometric Identity Management System Using Multiresolution Feature Learning

EIS-BIMS
Visual Abstract Image
Objectives
  1. To design a biometric authentication system using multiresolution signal processing–based machine learning techniques.
  2. To demonstrate data-driven computational efficiency by enabling system operation without model retraining or with minimal model adaptation.
  3. To establish interpretability of the proposed system through analysis and visualization of multiresolution feature maps.
  4. To demonstrate data-driven scalability and adaptability of the system for incorporating new users or data with minimal or no retraining overhead.
  5. To develop a mobile application for real-time biometric authentication using the proposed framework.
  6. To develop a biometrics dataset for cross-validation of hypotheisis.
  7. To check liveness of the biometrics sample.
  8. To identify AI generated fake biometrics samples.

 

Issues Involved

Identity management systems are widely deployed across contemporary applications, primarily to authenticate users based on their physiological characteristics. A significant proportion of existing systems rely on machine learning (ML) and deep learning (DL) frameworks. While these approaches have demonstrated strong recognition performance, they are often treated as black-box models with limited interpretability, weak mathematical transparency, and high dependence on large volumes of data. Moreover, such frameworks typically require substantial computational resources and frequent retraining to accommodate new users or data. Although advanced DL models exhibit adaptability, their performance often degrades when scaling to large populations, and their deployment becomes inefficient for small-scale systems where dense architectures lead to unnecessary resource consumption.

 

Team Lead

Prof. Divyang Jadav

Divyang.Jadav@nmims.edu


Team Members

  1. Name: Prof. Aditya Kasar

Email ID:Aditya.Kasar@nmims.edu

  1. Name: Prof. Tejaswini Chavan

Email ID:Tejaswini.Chavan@nmims.edu