Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

User Authentication and Vital Signs Extraction from Low-Frame-Rate and Monochrome No-contact Fingerprint Captures

Olaoluwayimika Olugbenle, Logan Drake, Naveenkumar G. Venkataswamy, Arfina Rahman, Yemi Afolayanka, Masudul Imtiaz, Mahesh K. Banavar

TL;DR

The paper tackles fingerprint authentication spoofing by leveraging PPG signals extracted from low-frame-rate blue-light fingertip videos to enable liveness detection and user identification. It presents a practical pipeline using a COTS no-contact sensor at 14 fps to derive PPG, estimate heart rate with a multi-filter approach, and validate a traditional peak-based Human ID as well as a CNN–LSTM deep learning method for user authentication. Key findings show an HR estimation mean error drop to 11.6% with enhanced filtering, and mixed identification results across users, with deep learning hindered by class imbalance. The work suggests that integrating PPG-based vital signs with fingerprint sensing can improve security and opens avenues for further optimization and broader datasets for robust biometric monitoring.

Abstract

We present our work on leveraging low-frame-rate monochrome (blue light) videos of fingertips, captured with an off-the-shelf fingerprint capture device, to extract vital signs and identify users. These videos utilize photoplethysmography (PPG), commonly used to measure vital signs like heart rate. While prior research predominantly utilizes high-frame-rate, multi-wavelength PPG sensors (e.g., infrared, red, or RGB), our preliminary findings demonstrate that both user identification and vital sign extraction are achievable with the low-frame-rate data we collected. Preliminary results are promising, with low error rates for both heart rate estimation and user authentication. These results indicate promise for effective biometric systems. We anticipate further optimization will enhance accuracy and advance healthcare and security.

User Authentication and Vital Signs Extraction from Low-Frame-Rate and Monochrome No-contact Fingerprint Captures

TL;DR

The paper tackles fingerprint authentication spoofing by leveraging PPG signals extracted from low-frame-rate blue-light fingertip videos to enable liveness detection and user identification. It presents a practical pipeline using a COTS no-contact sensor at 14 fps to derive PPG, estimate heart rate with a multi-filter approach, and validate a traditional peak-based Human ID as well as a CNN–LSTM deep learning method for user authentication. Key findings show an HR estimation mean error drop to 11.6% with enhanced filtering, and mixed identification results across users, with deep learning hindered by class imbalance. The work suggests that integrating PPG-based vital signs with fingerprint sensing can improve security and opens avenues for further optimization and broader datasets for robust biometric monitoring.

Abstract

We present our work on leveraging low-frame-rate monochrome (blue light) videos of fingertips, captured with an off-the-shelf fingerprint capture device, to extract vital signs and identify users. These videos utilize photoplethysmography (PPG), commonly used to measure vital signs like heart rate. While prior research predominantly utilizes high-frame-rate, multi-wavelength PPG sensors (e.g., infrared, red, or RGB), our preliminary findings demonstrate that both user identification and vital sign extraction are achievable with the low-frame-rate data we collected. Preliminary results are promising, with low error rates for both heart rate estimation and user authentication. These results indicate promise for effective biometric systems. We anticipate further optimization will enhance accuracy and advance healthcare and security.

Paper Structure

This paper contains 13 sections, 1 equation, 5 figures, 2 tables.

Table of Contents

  1. Introduction
  2. Fingerprint sensor and PPG extraction
  3. Vital Signs Extraction
  4. Improving Accuracy
  5. Results
  6. User Identification
  7. Human ID
  8. Algorithm
  9. Results
  10. Deep Learning-Based Method
  11. Algorithm
  12. Results
  13. Conclusions and Future Work

Figures (5)

  • Figure 1: a) Originally captured image. b) Image after cropping. We use an off-the-shelf no-contact fingerprint sensor that collects blue-light 3000ppi images at a low frame rate of 14 frames per second. The image obtained from the sensor is cropped to retain only the image of the fingertip and eliminate background artifacts.
  • Figure 2: Extracting the heart rate from the PPG. The algorithm is described in Section \ref{['sec:Vitals']}.
  • Figure 3: Grayscale PPG Pipeline.
  • Figure 4: Human ID Pipeline.
  • Figure 5: Deep-learning-based user identification pipeline from low frame-rate monochrome PPG signals. The system consists of two CNNs, two LSTMs, and a classification stage.