AI Paper Atlas

Research map / Video / Action recognition

Efficient Action Recognition via Dynamic Knowledge Propagation

Hanul Kim, Mihir Jain, Jun-Tae Lee, Sungrack Yun, Fatih Porikli

ICCV 2021 · Video › Action recognition · approach: Disentangled action recognition · 27 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
replace static averaging with dynamic knowledge propagation (ablation)student-teacher ensemble · ActivityNet 1.3mAP (%): 79.0 -> 81.2Improved
replace weighted-average fusion with dynamic knowledge propagation (ablation)student-teacher ensemble · ActivityNet 1.3mAP (%): 80.2 -> 81.2Improved
add knowledge distillation to student (ablation)Student · ActivityNet 1.3mAP (%): 77.6 -> 78.4Improved
add teacher on 5 frames at inference via dynamic propagation vs student with KD aloneStudent with KD · ActivityNet 1.3mAP (%): 78.4 -> 81.2Improved
increase teacher frames from 5 to 20Teacher · ActivityNet 1.3mAP (%): 80.1 -> 81.7Improved
add temporal feature fusion in both student and teacher (ablation)Dynamic-STE · ActivityNet 1.3mAP (%): 79.3 -> 81.2Improved
add temporal feature fusion in student only (ablation)Dynamic-STE · ActivityNet 1.3mAP (%): 79.3 -> 79.9Improved
add temporal feature fusion in teacher only (ablation)Dynamic-STE · ActivityNet 1.3mAP (%): 79.3 -> 80.7Improved
add frame classification loss (ablation)Dynamic-STE · ActivityNet 1.3mAP (%): 80.8 -> 81.0Improved
add cosine similarity loss (ablation)Dynamic-STE · ActivityNet 1.3mAP (%): 80.8 -> 80.9Improved
add both frame and cosine losses (ablation)Dynamic-STE · ActivityNet 1.3mAP (%): 80.8 -> 81.2Improved
Dynamic-STE with EfficientNet vs AR-Netefficient action recognition · ActivityNet 1.3mAP (%): 79.7 -> 81.2Improved

Earlier methods this paper compares with

Named baselines from earlier accepted papers in this paper's results tables: better than 3 of 3 on most shared results.

Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.

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