AI Paper Atlas

Research map / Detection and segmentation / Weakly supervised detection and segmentation

Tobias: A Random CNN Sees Objects

Yun-Hao Cao, Jianxin Wu

TPAMI 2023 · Detection and segmentation › Weakly supervised detection and segmentation · approach: Weakly supervised object localization · 4 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 ReLU instead of sigmoid (ablation)SCDA random R-50 · CUB-200 localizationaccuracy: 22.6 -> 41.8Improved
add network depth: bottlenecks [1,1,1,1] to [3,4,6,3] (ablation)SCDA random R-50 · ImageNet localizationaccuracy: 42.1 -> 48.2Improved
make network deeper: [3,4,6,3] to [6,8,12,6] (ablation)SCDA random R-50 · ImageNet localizationaccuracy: 48.2 -> 50.0Improved
add skip connections (ablation)random R-50 · ImageNet localizationaccuracy: 50.8 -> 48.2No clear effect
add batch normalization (ablation)random R-50 · ImageNet localizationaccuracy: 49.1 -> 48.2No clear effect
use Kaiming instead of Xavier initialization (ablation)random R-50 · ImageNet localizationaccuracy: 42.2 -> 48.2Improved
use CNN instead of ViT-Base for random-init localization (ablation)SCDA · ImageNet localizationaccuracy: 40.5 -> 48.2Improved
add Tobias background-swap augmentation (ablation)MoCov2 ResNet-50 CUB-200 200 ep · CUB-200 fine-tuningaccuracy: 56.2 -> 63.6Improved
add Tobias augmentation (ablation)MoCov2 ResNet-50 CUB-200 800 ep · CUB-200 fine-tuningaccuracy: 66.5 -> 67.2Improved
add Tobias augmentation (ablation)MoCov2 IN-10k 800 ep · VOC07+12 detectionAP75: 45.9 -> 48.5Improved
add Tobias augmentation (ablation)MoCov2 IN-10k 200 ep · Tiny-IN-200 linearaccuracy: 0.5 -> 9.9Improved
add Tobias augmentation (ablation)MoCov2 IN-50k 200 ep · Tiny-IN-200 linearaccuracy: 26.3 -> 26.0Hurt

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