Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Deep adaptative spectral zoom for improved remote heart rate estimation

Joaquim Comas, Adria Ruiz, Federico Sukno

TL;DR

The objective of the proposed model is to tailor the CZT to match the characteristics of each specific dataset sensor, facilitating a more optimal and accurate estimation of HR from the rPPG signal without compromising generalization across diverse datasets.

Abstract

Recent advances in remote heart rate measurement, motivated by data-driven approaches, have notably enhanced accuracy. However, these improvements primarily focus on recovering the rPPG signal, overlooking the implicit challenges of estimating the heart rate (HR) from the derived signal. While many methods employ the Fast Fourier Transform (FFT) for HR estimation, the performance of the FFT is inherently affected by a limited frequency resolution. In contrast, the Chirp-Z Transform (CZT), a generalization form of FFT, can refine the spectrum to the narrow-band range of interest for heart rate, providing improved frequential resolution and, consequently, more accurate estimation. This paper presents the advantages of employing the CZT for remote HR estimation and introduces a novel data-driven adaptive CZT estimator. The objective of our proposed model is to tailor the CZT to match the characteristics of each specific dataset sensor, facilitating a more optimal and accurate estimation of HR from the rPPG signal without compromising generalization across diverse datasets. This is achieved through a Sparse Matrix Optimization (SMO). We validate the effectiveness of our model through exhaustive evaluations on three publicly available datasets UCLA-rPPG, PURE, and UBFC-rPPG employing both intra- and cross-database performance metrics. The results reveal outstanding heart rate estimation capabilities, establishing the proposed approach as a robust and versatile estimator for any rPPG method.

Deep adaptative spectral zoom for improved remote heart rate estimation

TL;DR

The objective of the proposed model is to tailor the CZT to match the characteristics of each specific dataset sensor, facilitating a more optimal and accurate estimation of HR from the rPPG signal without compromising generalization across diverse datasets.

Abstract

Recent advances in remote heart rate measurement, motivated by data-driven approaches, have notably enhanced accuracy. However, these improvements primarily focus on recovering the rPPG signal, overlooking the implicit challenges of estimating the heart rate (HR) from the derived signal. While many methods employ the Fast Fourier Transform (FFT) for HR estimation, the performance of the FFT is inherently affected by a limited frequency resolution. In contrast, the Chirp-Z Transform (CZT), a generalization form of FFT, can refine the spectrum to the narrow-band range of interest for heart rate, providing improved frequential resolution and, consequently, more accurate estimation. This paper presents the advantages of employing the CZT for remote HR estimation and introduces a novel data-driven adaptive CZT estimator. The objective of our proposed model is to tailor the CZT to match the characteristics of each specific dataset sensor, facilitating a more optimal and accurate estimation of HR from the rPPG signal without compromising generalization across diverse datasets. This is achieved through a Sparse Matrix Optimization (SMO). We validate the effectiveness of our model through exhaustive evaluations on three publicly available datasets UCLA-rPPG, PURE, and UBFC-rPPG employing both intra- and cross-database performance metrics. The results reveal outstanding heart rate estimation capabilities, establishing the proposed approach as a robust and versatile estimator for any rPPG method.
Paper Structure (20 sections, 10 equations, 4 figures, 3 tables)

This paper contains 20 sections, 10 equations, 4 figures, 3 tables.

Table of Contents

  1. Introduction
  2. Contribution
  3. Related work
  4. rPPG measurement
  5. HR estimation from rPPG
  6. Methodology
  7. Chirp-Z Transform
  8. Deep CZT estimator
  9. Loss function
  10. Experiments
  11. Datasets
  12. Implementation details
  13. Preprocessing and training procedure
  14. Metrics and evaluation
  15. Preliminary experiment: Why CZT for remote HR estimation?
  16. ...and 5 more sections

Figures (4)

  • Figure 1: Frequency spectrum comparison between FFT (above) and CZT (below) for a UCLA-rPPG PPG signal sample (8.5 sec windows). In red is denoted the predicted HR while green represents the HR ground truth.
  • Figure 2: Overall structure of our proposed deep CZT adaptative estimator model.
  • Figure 3: MAE comparison between handcrafted HR estimation methods using different temporal windows with ground truth data.
  • Figure 4: Sparse Matrix optimization result. Difference between $\tilde{W}$ matrix of standard CZT and $\tilde{W}'$ matrix of trained deep CZT estimator.