AI Paper Atlas

Research map / Detection and segmentation / CNN features

Detecting Rotated Objects as Gaussian Distributions and Its 3-D Generalization

Xue Yang, Gefan Zhang, Xiaojiang Yang, Yue Zhou, Wentao Wang, Jin Tang, Tao He, Junchi Yan

TPAMI 2022 · Detection and segmentation › CNN features · approach: Box regression and loss design · 112 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
replace Smooth L1 with GWD loss (D_oc)RetinaNet · DOTA-v1.0mAP50: 65.73 -> 68.93Improved
replace Smooth L1 with KLD loss (D_oc)RetinaNet · DOTA-v1.0mAP50: 65.73 -> 71.28Improved
replace Smooth L1 with BCD loss (D_oc)RetinaNet · DOTA-v1.0mAP50: 65.73 -> 71.23Improved
replace Smooth L1 with KLD loss (D_le)RetinaNet · DOTA-v1.0mAP50: 64.17 -> 68.88Improved
replace Smooth L1 with KLD lossRetinaNet · HRSC2016AP50:95: 47.76 -> 57.80Improved
replace Smooth L1 with KLD lossR3Det · HRSC2016AP75: 43.42 -> 77.38Improved
replace Smooth L1 with KLD lossRetinaNet · FDDBAP85: 12.67 -> 42.20Improved
replace Smooth L1 with GWD lossRetinaNet · MSRA-TD500Hmean50: 70.98 -> 76.76Improved
replace Smooth L1 with BCD lossRetinaNet · ICDAR2015 (R+F)Hmean85: 11.59 -> 10.53Hurt
replace Smooth L1 with KLD lossPointPillars · KITTI valmAP moderate: 64.28 -> 66.19Improved
replace Smooth L1 with GWD lossPointPillars · KITTI val car 3-DEasy AP: 88.26 -> 87.38Hurt
add normalization by Eq. to Smooth L1 (ablation)RetinaNet · HRSC2016Hmean50:95: 47.76 -> 43.47Hurt

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