AI Paper Atlas

Research map / Detection and segmentation / CNN features

Estimating and Evaluating Regression Predictive Uncertainty in Deep Object Detectors

Ali Harakeh, Steven L. Waslander

ICLR 2021 · Detection and segmentation › CNN features · approach: Region-based and query-based detectors · 24 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
Train DETR bounding box distributions with the energy score instead of NLL (ablation)DETR · COCOmAP (%): 37.10 -> 39.96Improved
Train RetinaNet bounding box distributions with the energy score instead of NLL (ablation)RetinaNet · COCOmAP (%): 35.11 -> 36.02Improved
Train FasterRCNN bounding box distributions with the energy score instead of NLL (ablation)FasterRCNN · COCOmAP (%): 37.14 -> 37.32Improved
Train DETR bounding box distributions with the energy score instead of DMM (ablation)DETR · COCOmAP (%): 39.82 -> 39.96Improved
Train RetinaNet bounding box distributions with DMM instead of the energy score (ablation)RetinaNet · COCOmAP (%): 36.02 -> 36.28Improved
Train FasterRCNN bounding box distributions with the energy score instead of DMM (ablation)FasterRCNN · COCOmAP (%): 37.24 -> 37.32Improved
Train DETR bounding box distributions with the energy score instead of NLL (ablation)DETR · COCOCE (Reg): 0.0512 -> 0.0068Improved
Train RetinaNet bounding box distributions with the energy score instead of NLL (ablation)RetinaNet · COCOCE (Reg): 0.0242 -> 0.0059Improved
Train FasterRCNN bounding box distributions with the energy score instead of NLL (ablation)FasterRCNN · COCOCE (Reg): 0.0049 -> 0.0051Hurt
Train DETR or RetinaNet bounding box distributions with NLL instead of the energy score (ablation)DETR and RetinaNet · COCO validationPredictive entropyHurt
Use ad-hoc IOU-threshold regression target assignment rather than broad target supportRetinaNet and FasterRCNN variance networks · COCO validationUncertainty reliability across prediction errorsHurt

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