# Explaining RhythmFormer: A Systematic XAI Analysis of Periodic Sparse Attention for Remote Photoplethysmography

arXiv: https://arxiv.org/abs/2606.13839 · HTML: https://mlsanity.com/paper/2606.13839/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Builds on 28 of its 41 references that are in the map.
- First posted on arXiv 11 Jun 2026.

## Builds on (most cited first, 28 of 28)

- [Attention Is All You Need](https://mlsanity.com/paper/1706.03762/) (2017)
- [An Image is Worth 16x16 Words: Transformers for Image Recognition at Scale](https://mlsanity.com/paper/2010.11929/) (2020)
- [Grad-CAM: Visual Explanations from Deep Networks via Gradient-based Localization](https://mlsanity.com/paper/1610.02391/) (2016)
- [A Unified Approach to Interpreting Model Predictions](https://mlsanity.com/paper/1705.07874/) (2017)
- ["Why Should I Trust You?": Explaining the Predictions of Any Classifier](https://mlsanity.com/paper/1602.04938/) (2016)
- [Axiomatic Attribution for Deep Networks](https://mlsanity.com/paper/1703.01365/) (2017)
- [Deep Inside Convolutional Networks: Visualising Image Classification Models and Saliency Maps](https://mlsanity.com/paper/1312.6034/) (2013)
- [Attention is not Explanation](https://mlsanity.com/paper/1902.10186/) (2019)
- [Quantifying Attention Flow in Transformers](https://mlsanity.com/paper/2005.00928/) (2020)
- [BiSeNet: Bilateral Segmentation Network for Real-time Semantic Segmentation](https://mlsanity.com/paper/1808.00897/) (2018)
- [Attention is not not Explanation](https://mlsanity.com/paper/1908.04626/) (2019)
- [Is Attention Interpretable?](https://mlsanity.com/paper/1906.03731/) (2019)
- [DeepPhys: Video-Based Physiological Measurement Using Convolutional Attention Networks](https://mlsanity.com/paper/1805.07888/) (2018)
- [BiFormer: Vision Transformer with Bi-Level Routing Attention](https://mlsanity.com/paper/2303.08810/) (2023)
- [PhysFormer: Facial Video-based Physiological Measurement with Temporal Difference Transformer](https://mlsanity.com/paper/2111.12082/) (2021)
- [Learning to Deceive with Attention-Based Explanations](https://mlsanity.com/paper/1909.07913/) (2019)
- [Remote Heart Rate Measurement from Highly Compressed Facial Videos: an End-to-end Deep Learning Solution with Video Enhancement](https://mlsanity.com/paper/1907.11921/) (2019)
- [EfficientPhys: Enabling Simple, Fast and Accurate Camera-Based Vitals Measurement](https://mlsanity.com/paper/2110.04447/) (2021)
- [VIPL-HR: A Multi-modal Database for Pulse Estimation from Less-constrained Face Video](https://mlsanity.com/paper/1810.04927/) (2018)
- [MMPD: Multi-Domain Mobile Video Physiology Dataset](https://mlsanity.com/paper/2302.03840/) (2023)
- [A Reproducible Study on Remote Heart Rate Measurement](https://mlsanity.com/paper/1709.00962/) (2017)
- [PhysFormer++: Facial Video-based Physiological Measurement with SlowFast Temporal Difference Transformer](https://mlsanity.com/paper/2302.03548/) (2023)
- [Explainable Artificial Intelligence: A Survey of Needs, Techniques, Applications, and Future Direction](https://mlsanity.com/paper/2409.00265/) (2024)
- [RhythmFormer: Extracting Patterned rPPG Signals based on Periodic Sparse Attention](https://mlsanity.com/paper/2402.12788/) (2024)
- [On the Faithfulness of Vision Transformer Explanations](https://mlsanity.com/paper/2404.01415/) (2024)
- [Dual-path TokenLearner for Remote Photoplethysmography-based Physiological Measurement with Facial Videos](https://mlsanity.com/paper/2308.07771/) (2023)
- [CodePhys: Robust Video-based Remote Physiological Measurement through Latent Codebook Querying](https://mlsanity.com/paper/2502.07526/) (2025)
- [DD-rPPGNet: De-interfering and Descriptive Feature Learning for Unsupervised rPPG Estimation](https://mlsanity.com/paper/2407.21402/) (2024)

## Abstract (arXiv)

Remote photoplethysmography (rPPG) transformers achieve low heart-rate error on benchmarks, yet their decisions remain opaque--a growing concern as rPPG moves toward clinical heart rate estimation. Existing rPPG XAI is dominated by qualitative heatmap inspection without quantitative faithfulness metrics or physiology-grounded validation, leaving a gap between visual plausibility and auditable evidence. We address this gap. First, we adapt four attribution methods (raw attention, rollout, flow, Beyond Intuition) to RhythmFormer's bi-level routing attention with top-k selection. Second, we introduce a skin coverage metric quantifying how much attribution mass falls on skin regions. Third, we adapt the SaCo faithfulness coefficient from its original classification setting to rPPG regression by using the MAE between original and perturbed predicted rPPG waveforms as the perturbation impact. Applying these tools, we quantify a multi-hop leakage effect under sparse top-k routing: attention rollout and flow almost completely restores the connections that individual refined-attention layers explicitly set to zero. Beyond Intuition mitigates this via its value-projection-weighted rollout and gradient-supported mask, attaining the highest median refined skin coverage (0.83 vs. 0.57 for vanilla rollout) and faithfulness (F=0.92) among the evaluated methods on UBFC-rPPG. Validation across diverse datasets and model variants is needed. A case study on a low-SaCo outlier further shows all four methods recovering consistently once an artefactual region is replaced, suggesting consistent SaCo behavior across attribution families in this illustrative case. Together, these metrics move XAI for rPPG toward auditable numerical evidence about spatial alignment and perturbation faithfulness, i.e. trustworthy rPPG XAI.