AI Paper Atlas

Research map / Continual learning / Online and replay-based

Continual Learning for Remote Physiological Measurement: Minimize Forgetting and Simplify Inference

Qian Liang, Yan Chen, Yang Hu

ECCV 2024 · Continual learning › Online and replay-based · approach: Unified and unsupervised continual frameworks · 6 citations (OpenAlex)

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
add style prototype augmentation to adapter finetuning (ablation)ADDP · rPPG DIL protocolMAE_N: 4.59 -> 4.23Improved
add noise prototype augmentation on top of style augmentation (ablation)ADDP · rPPG DIL protocolMAE_N: 4.23 -> 4.10Improved
add style augmentation (before noise and inference simplification) (ablation)ADDP · rPPG DIL protocolR_N: 0.78 -> 0.73Hurt
add prototype-based inference simplification (ablation)ADDP · rPPG DIL protocolMAE_N: 4.10 -> 3.70Improved
use test-time style shifting instead of inference simplification (ablation)ADDP · rPPG DIL protocolMAE_N: 3.70 -> 3.91Hurt
add DiffNorm module to Uniformer (ablation)Uniformer · VIPL Fold1-4MAE: 5.56 -> 4.84Improved
use EfficientPhys normalization instead of DiffNorm on Uniformer (ablation)Uniformer · MMPDMAE: 6.05 -> 7.15Hurt
use vanilla ViT backbone instead of Uniformer (ablation)rPPG backbone · VIPL Fold1-4MAE: 5.56 -> 12.52Hurt
use LAE offline adapter or ensemble in rPPG continual learningLAE · rPPG DIL protocolMAE_N: 7.94 (offline), 8.07 (ensemble) vs 4.70 (online)Hurt

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy