AI Paper Atlas

Research map / Detection and segmentation / CNN features

Gaussian YOLOv3: An Accurate and Fast Object Detector Using Localization Uncertainty for Autonomous Driving

Jiwoong Choi, Dayoung Chun, Hyun Kim, Hyuk-Jae Lee

ICCV 2019 · Detection and segmentation › CNN features · approach: Box regression and loss design · 520 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 Gaussian bbox modeling, NLL loss and uncertainty-aware detection criterionYOLOv3 (512x512) · KITTI valmAP: 80.52 -> 83.61Improved
add Gaussian bbox modeling, NLL loss and uncertainty-aware detection criterionYOLOv3 (512x512) · BDD testmAP: 14.9 -> 18.4Improved
add Gaussian modeling and uncertainty criterionYOLOv3 · COCOAP: 36.1, 3.1 higher than YOLOv3Improved
add Gaussian modeling and uncertainty criterionYOLOv3 · COCOAP75: 39.0, 4.6 higher than YOLOv3Improved
add Gaussian modeling and uncertainty criterionYOLOv3 · KITTI valfalse positives: 1,681 -> 985 (-41.40%)Improved
add Gaussian modeling and uncertainty criterionYOLOv3 · KITTI valtrue positives: 13,575 -> 14,560 (+7.26%)Improved
add Gaussian modeling and uncertainty criterionYOLOv3 · BDD valfalse positives: 86,380 -> 51,296 (-40.62%)Improved
add Gaussian modeling and uncertainty criterionYOLOv3 · BDD valtrue positives: 57,261 -> 59,724 (+4.30%)Improved
add Gaussian modeling and uncertainty criterionYOLOv3 (512x512) · KITTI valFPS: 43.57 -> 43.13No clear effect
add Gaussian modeling and uncertainty criterionYOLOv3 · BDD testFPS: 42.9 -> 42.5No clear effect
raise input resolution from 512x512 to 704x704Gaussian YOLOv3 · KITTI valmAP: 83.61 -> 86.79 mAP, 43.13 -> 24.91 FPSMixed
raise input resolution from 512x512 to 736x736Gaussian YOLOv3 · BDD testmAP: 18.4 -> 20.8 mAP, 42.5 -> 22.5 FPSMixed

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