AI Paper Atlas

Research map / Video / Action recognition

MammalNet: A Large-Scale Video Benchmark for Mammal Recognition and Behavior Understanding

Jun Chen, Ming Hu, Darren J. Coker, Michael L. Berumen, Blair Costelloe, Sara Beery, Anna Rohrbach, Mohamed Elhoseiny

CVPR 2023 · Video › Action recognition · approach: Disentangled action recognition · 42 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
initialize MViT V2 with Kinetics 400 pretrainingMViT V2 · MammalNet standard classificationanimal per-class top-1 accuracy: 19.7 to 32.5Improved
initialize MViT V2 with Kinetics 400 pretrainingMViT V2 · MammalNet standard classificationbehavior per-class top-1 accuracy: 27.7 to 37.8Improved
add one labeled example per animal-behavior compositionMViT V2* · MammalNet compositional low-shotbehavior per-example accuracy: 26.2 to 28.9Improved
add five labeled examples per animal-behavior composition rather than zeroMViT V2* · MammalNet compositional low-shotbehavior per-example accuracy: 26.2 to 31.7Improved
add five labeled examples per animal-behavior composition rather than zeroMViT V2* · MammalNet compositional low-shotanimal per-example accuracy: 32.2 to 39.3Improved
train animal and behavior recognition jointly instead of separatelyMammalNet classification model · MammalNet standard classificationbehavior per-class accuracy: ~2.2 per-class accuracyImproved
use joint animal-and-behavior training rather than animal-only predictionMammalNet classification model · MammalNet standard classificationanimal recognitionHurt
transfer behavior knowledge from seen animals to unseen animal-behavior compositionsMViT V2* · MammalNet compositional low-shot (0-shot)behavior per-example accuracy: 26.2Improved

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