rFaceNet: An End-to-End Network for Enhanced Physiological Signal Extraction through Identity-Specific Facial Contours

TL;DR

rFaceNet is a novel end-to-end neural network that improves physiological signal extraction from facial videos by leveraging identity-specific contours and advanced feature integration, outperforming existing methods in heart rate estimation.

Contribution

The paper introduces rFaceNet, a new deep learning framework that enhances rPPG signal extraction by incorporating facial contours and specialized modules for better focus and interpretability.

Findings

Significantly improves heart rate estimation accuracy.

Outperforms state-of-the-art methods on benchmark datasets.

Enhances interpretability of physiological signals.

Abstract

Remote photoplethysmography (rPPG) technique extracts blood volume pulse (BVP) signals from subtle pixel changes in video frames. This study introduces rFaceNet, an advanced rPPG method that enhances the extraction of facial BVP signals with a focus on facial contours. rFaceNet integrates identity-specific facial contour information and eliminates redundant data. It efficiently extracts facial contours from temporally normalized frame inputs through a Temporal Compressor Unit (TCU) and steers the model focus to relevant facial regions by using the Cross-Task Feature Combiner (CTFC). Through elaborate training, the quality and interpretability of facial physiological signals extracted by rFaceNet are greatly improved compared to previous methods. Moreover, our novel approach demonstrates superior performance than SOTA methods in various heart rate estimation benchmarks.