CM-SSL (cross-modal token-vocabulary SSL for ECG/PPG)
UnconfirmedCatalog paper date: Not reportedAdded to atlas: Oct 11, 2026
Activity counts measure attention, not model quality. Missing counts are unknown. Publication status follows the linked source; a listed venue alone does not confirm peer review.
CM-SSL trains single-signal ECG or PPG encoders against a shared token vocabulary learned from synchronized ECG and PPG pairs. A joint tokenizer is pretrained first, then the unimodal encoder, then the model is fine-tuned. Pretraining uses PulseDB, and fine-tuning covers UCI BP, BIDMC and PPG-DaLiA for PPG and CSN, PhysioNet 2017 and PTB-XL for ECG.
- Method: cross-modal joint tokenizer + unimodal encoder
- Pretraining data: PulseDB
- Fine-tuning: UCI BP, BIDMC, PPG-DaLiA, CSN, PhysioNet 2017, PTB-XL
- Code: GitHub (ximosss/CM-SSL)
At a glance
- Type
- Foundation model
- Form factor
- Not reported
- Parameters
- Not reported
- Architecture
- Not reported
- Venue
- Not reported
- Year
- Not reported
- arXiv
- Not reported
- DOI
- Not reported
- Weights
- Not released
- Licence
- Not reported
- Datasets
- Not reported
- Added
- Oct 11, 2026
- Updated
- Oct 11, 2026
How it works
- Inputs
- PPG, ECG
- Approach
- PPG · ECG · SSL · tokenizer · icassp 2026
Datasets
No public datasets linked yet.
Papers· 1
Cite
@misc{cmssl,
title = {{CM-SSL (cross-modal token-vocabulary SSL for ECG/PPG)}}
}History
Added to the ChatPPG research hub