# Template Collapse and Information-Theoretic Limits in Camera rPPG Pulse Morphology Restoration

arXiv: https://arxiv.org/abs/2606.03802 · HTML: https://mlsanity.com/paper/2606.03802/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Builds on 23 of its 41 references that are in the map.
- First posted on arXiv 2 Jun 2026.

## Builds on (most cited first, 23 of 23)

- [Supervised Contrastive Learning](https://mlsanity.com/paper/2004.11362/) (2020)
- [VICReg: Variance-Invariance-Covariance Regularization for Self-Supervised Learning](https://mlsanity.com/paper/2105.04906/) (2021)
- [Understanding Dimensional Collapse in Contrastive Self-supervised Learning](https://mlsanity.com/paper/2110.09348/) (2021)
- [Toward Multimodal Image-to-Image Translation](https://mlsanity.com/paper/1711.11586/) (2017)
- [Cross Modal Distillation for Supervision Transfer](https://mlsanity.com/paper/1507.00448/) (2015)
- [Joint Generative and Contrastive Learning for Unsupervised Person Re-identification](https://mlsanity.com/paper/2012.09071/) (2020)
- [Joint Contrastive Learning with Infinite Possibilities](https://mlsanity.com/paper/2009.14776/) (2020)
- [Motion-Focused Contrastive Learning of Video Representations](https://mlsanity.com/paper/2201.04029/) (2022)
- [Deep Latent Space Learning for Cross-modal Mapping of Audio and Visual Signals](https://mlsanity.com/paper/1909.08685/) (2019)
- [Towards Fine-grained Visual Representations by Combining Contrastive Learning with Image Reconstruction and Attention-weighted Pooling](https://mlsanity.com/paper/2104.04323/) (2021)
- [Guided Variational Autoencoder for Speech Enhancement With a Supervised Classifier](https://mlsanity.com/paper/2102.06454/) (2021)
- [How Suboptimal is Training rPPG Models with Videos and Targets from Different Body Sites?](https://mlsanity.com/paper/2403.10582/) (2024)
- [Self-supervised contrastive learning performs non-linear system identification](https://mlsanity.com/paper/2410.14673/) (2024)
- [SIGMA-PPG: Statistical-prior Informed Generative Masking Architecture for PPG Foundation Model](https://mlsanity.com/paper/2601.21031/) (2026)
- [PPGFlowECG: Latent Rectified Flow with Cross-Modal Encoding for PPG-Guided ECG Generation and Cardiovascular Disease Detection](https://mlsanity.com/paper/2509.19774/) (2025)
- [Cross-modal Variational Auto-encoder with Distributed Latent Spaces and Associators](https://mlsanity.com/paper/1905.12867/) (2019)
- [Variational Supervised Contrastive Learning](https://mlsanity.com/paper/2506.07413/) (2025)
- [FreqPhys: Repurposing Implicit Physiological Frequency Prior for Robust Remote Photoplethysmography](https://mlsanity.com/paper/2604.00534/) (2026)
- [Reduction of motion artifacts from photoplethysmography signals using learned convolutional sparse coding](https://mlsanity.com/paper/2508.10805/) (2025)
- [SyncGAN: Synchronize the Latent Space of Cross-modal Generative Adversarial Networks](https://mlsanity.com/paper/1804.00410/) (2018)
- [Weighted Combination and Singular Spectrum Analysis Based Remote Photoplethysmography Pulse Extraction in Low-light Environments](https://mlsanity.com/paper/2503.03780/) (2025)
- [SkinMap: Weighted Full-Body Skin Segmentation for Robust Remote Photoplethysmography](https://mlsanity.com/paper/2510.05296/) (2025)
- [ToTMNet: FFT-Accelerated Toeplitz Temporal Mixing Network for Lightweight Remote Photoplethysmography](https://mlsanity.com/paper/2601.04159/) (2026)

## Abstract (arXiv)

Objective: Consumer face camera remote photoplethysmography (rPPG) enables passive cardiovascular monitoring, but whether single-cycle waveform morphology encoding arterial stiffness biomarkers is recoverable from this measurement has not been characterised. Methods: We evaluated 16 architectures spanning six families on 153 subjects across three datasets, introducing cross-subject Pearson r to distinguish subject-specific recovery from template collapse. Results: No architecture recovered subject-specific morphology (cross-subject r range 0.773--0.9999; ground-truth ceiling 0.601). Supervised Contrastive (SupCon) converged to log N = 4.844, constituting the strongest available empirical evidence that no discriminative morphological structure is extractable from single-cycle rPPG by the encoder families tested. The VAE decoder restores population-level harmonic content absent from the rPPG input (H2/H1: 0.310 output vs. 0.275 input), generalising zero-shot to UBFC (r = +0.708); a directional hallucination gap (p = 0.150) suggests partial signal reading. Anti-collapse objectives fail when input carries no discriminative structure. Significance: Consumer cameras cannot encode individual arterial morphology; cross-subject r is a necessary collapse diagnostic for waveform reconstruction benchmarks.