AI Paper Atlas

Research map / Video / Action recognition

Large Scale Holistic Video Understanding

Ali Diba, Mohsen Fayyaz, Vivek Sharma, Manohar Paluri, Jürgen Gall, Rainer Stiefelhagen, Luc Van Gool

ECCV 2020 · Video › Action recognition · approach: Contrastive video representation learning · 121 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 a 2D appearance branch to the 3D network (HATNet)3D-ResNet18 · HVU overallmAP: 35.8 -> 40Improved
add the 2D appearance branch to the 3D-STCNet3D-STCNet · HVU overallmAP: 36.7 -> 40Improved
train with six task-specific heads instead of single standard head3D-ResNet18 · HVU overallmAP: 35.8 -> 37Improved
train with six task-specific heads instead of single standard headHATNet · HVU overallmAP: 40 -> 41.3Improved
pre-train on HVU instead of Kinetics3D-ResNet18 · UCF101mAP: 89.8 -> 90.5Improved
pre-train on HVU instead of Kinetics3D-ResNet18 · HMDB51mAP: 62.1 -> 65.1Improved
pre-train on HVU instead of training from scratch3D-ResNet18 · KineticsmAP: 65.6 -> 67.8Improved
add object tags to action-only training (ablation)3D-ResNet18 · HVU action categoryaccuracy: 43.6 -> 44.5Improved
add scene tags to action+object training (ablation)3D-ResNet18 · HVU action categoryaccuracy: 44.5 -> 45.6Improved
train all categories together instead of action only (ablation)3D-ResNet18 · HVU action categoryaccuracy: 43.6 -> 47.5Improved
pre-train on HVU instead of Kinetics for captioning3D-ResNet50+M · MSR-VTTBLEU@4: 41.8 -> 42.7Improved
pre-train on HVU instead of Kinetics for clustering features3D-ResNet50 · ActivityNet-100clustering accuracy: 50.3 -> 53.5Improved

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