AI Paper Atlas

Research map / Video / Video-language understanding

The "Something Something" Video Database for Learning and Evaluating Visual Common Sense

Raghav Goyal, Samira Ebrahimi Kahou, Vincent Michalski, Joanna Materzynska, Susanne Westphal, Heuna Kim, Valentin Haenel, Ingo Fruend, Peter Yianilos, Moritz Mueller-Freitag, Florian Hoppe, Christian Thurau and 2 more

ICCV 2017 · Video › Video-language understanding · approach: Event and anomaly reasoning in video · 1,590 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 2D CNN with ImageNet pretrainingVGG-16 + Avg · Something-Something (10 classes)top-1 error (%): 76.5 -> 54.7Improved
initialize 2D CNN with ImageNet pretrainingVGG-16 + Avg · Something-Something (40 classes)top-1 error (%): 88.0 -> 79.2Improved
replace frame averaging with LSTMPre-2D CNN · Something-Something (10 classes)top-1 error (%): 54.7 -> 52.3Improved
replace frame averaging with LSTMPre-2D CNN · Something-Something (40 classes)top-1 error (%): 79.2 -> 77.8Improved
use 3D CNN instead of 2D CNN, both from scratchscratch baselines · Something-Something (40 classes)top-1 error (%): 88.0 -> 70.3Improved
use Sports-1M pretrained 3D CNN instead of pretrained 2D CNNpretrained baselines · Something-Something (10 classes)top-1 error (%): 54.7 -> 47.5Improved
combine 2D and 3D CNN encodingsbest 3D CNN · Something-Something (10 classes)top-1 error (%): 47.5 -> 44.9Improved
combine 2D and 3D CNN encodingsbest 3D CNN · Something-Something (40 classes)top-1 error (%): 66.2 -> 63.8Improved
increase number of classes from 10 to 40 to 174Pre-3D CNN + Avg · Something-Somethingtop-1 error (%): 47.5 -> 66.2 -> 88.5Hurt
add dropout on first FC layer and batch norm on last layer (ablation)baseline models · Something-SomethingtrainingImproved
add flipping and small random rotation augmentation (ablation)baseline models · Something-SomethingaccuracyNo clear effect

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