AI Paper Atlas

Research map / Detection and segmentation / CNN features

Gradient Harmonized Single-Stage Detector

Buyu Li, Yu Liu, Xiaogang Wang

AAAI 2019 · Detection and segmentation › CNN features · approach: Efficient DETR query design · 597 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 cross-entropy with GHM-C for classificationRetinaNet-ResNet-50 · COCO minivalAP: 28.6 -> 35.8Improved
Replace Focal Loss with GHM-C for classificationRetinaNet-ResNet-50 · COCO minivalAP: 35.6 -> 35.8Improved
Use GHM-C + GHM-R instead of Focal LossRetinaNet-FPN-ResNeXt-101 · COCO test-devAP: 40.8 -> 41.6Improved
Add GHM-R for box regression instead of Smooth L1RetinaNet-ResNet-50 with GHM-C · COCO minivalAP: 35.8 -> 36.4Improved
Add GHM-R for box regression instead of Smooth L1RetinaNet-ResNet-50 with GHM-C · COCO minivalAP@IoU=0.5: 55.5 -> 54.6Hurt
Add GHM-R for box regression instead of Smooth L1RetinaNet-ResNet-50 with GHM-C · COCO minivalAP: 36.4 -> 37.4Improved
Replace Smooth L1 with Authentic Smooth L1 (ablation)RetinaNet-ResNet-50 with GHM-C · COCO minivalAP: 35.8 -> 35.7No clear effect
Use 30 rather than 5 unit regions for GHM-C (ablation)RetinaNet-ResNet-50 · COCO minivalAP: 33.4 -> 35.8Improved
Increase GHM-C unit regions from 30 to 40 (ablation)RetinaNet-ResNet-50 · COCO minivalAP: 35.8 -> 35.4Hurt
Use unit-region approximation instead of the original gradient-density calculation (ablation)GHM-C · COCO minivalaverage time per iteration: 13.675 -> 0.824Improved
Use unit-region approximation instead of the original gradient-density calculation (ablation)GHM-C · COCO minivalAP: 35.9 -> 35.8No clear effect

Earlier methods this paper compares with

Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.

Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.

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