AI Paper Atlas

Research map / Label-efficient and robust learning / Domain adaptation and generalization

Bi-Level Alignment for Cross-Domain Crowd Counting

Shenjian Gong, Shanshan Zhang, Jian Yang, Dengxin Dai, Bernt Schiele

CVPR 2022 · Label-efficient and robust learning › Domain adaptation and generalization · approach: Feature alignment for unsupervised adaptation · 40 citations (OpenAlex)

Check 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
apply bi-level alignment (ablation)BLA with VGG16 · GCC → SHAMAE: 134.7 -> 99.3Improved
add task-driven data alignment (ablation)BLA without fine-grained feature alignment · GCC → SHAMAE: 134.7 -> 109.1Improved
replace global feature alignment with fine-grained feature alignment (ablation)BLA without task-driven data alignment · GCC → SHAMAE: 134.7 -> 121.1Improved
replace task-driven data alignment with style transfer (ablation)BLA · GCC → SHAMAE: 99.3 -> 119.4Hurt
replace task-driven data alignment with domain randomization (ablation)BLA · GCC → SHAMAE: 99.3 -> 110.0Hurt
use grid size (16,16) instead of the other tested grid sizes (ablation)fine-grained feature alignment · GCC → SHAMAE: (2,2): 138.1; (4,4): 130.5; (8,8): 129.0; (16,16): 121.1; (32,32): 141.4Improved
use the target dataset's searched transform rather than a transform searched for another dataset (ablation)BLA task-driven data alignment · GCC → SHA and GCC → SHBMAE: SHA transform: SHA 99.3, SHB 14.7; SHB transform: SHA 104.3, SHB 11.9Improved
use BLA instead of the strongest reported baselineVGG16 crowd counting · GCC → SHAMAE: 112.4 -> 99.3Improved
use BLA instead of the strongest reported baselineVGG16 crowd counting · GCC → SHBMAE: 12.8 -> 11.9Improved
use BLA instead of the strongest reported baselineVGG16 crowd counting · GCC → QNRFMAE: 210.0 -> 198.9Improved
use BLA instead of the strongest reported baselineVGG16 crowd counting · GCC → UCF-CC-50MAE: 355.0 -> 346.8Improved
use BLA instead of the strongest reported baselineVGG16 crowd counting · GCC → WorldExpo'10Avg. MAE: 17.4 -> 17.9Hurt

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