AI Paper Atlas

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

Self-Adversarial Disentangling for Specific Domain Adaptation

Qianyu Zhou, Qiqi Gu, Jiangmiao Pang, Xuequan Lu, Lizhuang Ma

TPAMI 2023 · Label-efficient and robust learning › Domain adaptation and generalization · approach: Adversarial domain adaptation · 67 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 Domainness Creator (ablation)GPA · Cityscapes to Foggy Cityscapes (VGG16)mAP: 39.5 -> 42.5Improved
add Self-Adversarial Regularizer after DC (ablation)GPA + DC · Cityscapes to Foggy CityscapesmAP: 42.5 -> 45.2Improved
add DC to AdaptSegNet (ablation)AdaptSegNet · Foggy Zurich++ segmentation (mIoU)mAP: 29.4 -> 33.5Improved
add SAR after DC (ablation)AdaptSegNet + DC · Foggy Zurich++ segmentation (mIoU)mAP: 33.5 -> 35.2Improved
use DC instead of random crop augmentation (ablation)GPA · Cityscapes to Foggy CityscapesmAP: 42.7 -> 45.2Improved
use DC instead of color jittering augmentation (ablation)GPA · Cityscapes to Foggy CityscapesmAP: 41.8 -> 45.2Improved
add L_inv alongside L_spf (ablation)GPA + DC · Cityscapes to Foggy CityscapesmAP: 43.1 -> 45.2Improved
add L_spf alongside L_inv (ablation)GPA + DC · Cityscapes to Foggy CityscapesmAP: 43.3 -> 45.2Improved
use SAR instead of MRL (ablation)AdaptSegNet + DC · Foggy Zurich++ segmentation (mIoU)mAP: 34.0 -> 35.2Improved
use SAR instead of CIDA (ablation)AdaptSegNet + DC · Foggy Zurich++ segmentation (mIoU)mAP: 34.2 -> 35.2Improved
increase domainness number N from 0 to 4 (ablation)Ours · Virtual KITTI to CKITTImAP: 46.14 -> 50.57Improved
increase domainness number N from 0 to 4 (ablation)Ours · Cityscapes to RainCityscapesmAP: 34.96 -> 44.21Improved

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