AI Paper Atlas

Research map / Video / Action recognition

CLIP-guided Prototype Modulating for Few-shot Action Recognition

Xiang Wang, Shiwei Zhang, Jun Cen, Changxin Gao, Yingya Zhang, Deli Zhao, Nong Sang

IJCV 2023 · Video › Action recognition · approach: Disentangled action recognition · 99 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 video-text contrastive objective (ablation)OTAM baseline with CLIP-RN50 · Kinetics 5-way 1-shotaccuracy (%): 76.4 -> 80.4Improved
add video-text contrastive objective (ablation)OTAM baseline with CLIP-RN50 · Kinetics 5-way 5-shotaccuracy (%): 88.9 -> 91.6Improved
add prototype modulation (ablation)OTAM baseline with CLIP-RN50 · SSv2-Small 5-way 1-shotaccuracy (%): 39.9 -> 50.8Improved
add prototype modulation (ablation)OTAM baseline with CLIP-RN50 · SSv2-Small 5-way 5-shotaccuracy (%): 51.9 -> 55.3Improved
add video-text objective on top of prototype modulation (ablation)CLIP-FSAR · SSv2-Small 5-way 1-shotaccuracy (%): 50.8 -> 52.0Improved
add CLIP text features to temporal Transformer modulation (ablation)Temporal Transformer-only · Kinetics 5-way 1-shotaccuracy (%): 77.9 -> 87.6Improved
use temporal Transformer instead of Bi-GRU (ablation)CLIP-FSAR modulation · SSv2-Small 5-way 1-shotaccuracy (%): 47.1 -> 52.0Improved
share Transformer weights between support and query (ablation)CLIP-FSAR · Kinetics 5-way 5-shotaccuracy (%): 84.7 -> 91.9Improved
increase temporal Transformer layers from 1 to 5 (ablation)CLIP-FSAR · Kinetics 5-way 1-shotaccuracy (%): 87.6 -> 85.2Hurt
increase temporal Transformer layers from 1 to 3 (ablation)CLIP-FSAR · SSv2-Small 5-way 1-shotaccuracy (%): 52.0 -> 52.4No clear effect
replace CLIP with ImageNet RN18 encoder plus CLIP-FSAROTAM · SSv2-Small 5-way 1-shotaccuracy (%): 35.8 -> 46.0Improved
plug CLIP-FSAR into TRX with CLIP-RN50TRX · Kinetics 5-way 1-shotaccuracy (%): 71.8 -> 86.7Improved

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