atlas

Research map / Detection and segmentation / CNN features

NAS-FPN: Learning Scalable Feature Pyramid Architecture for Object Detection

CVPR 2019 · Detection and segmentation › CNN features · approach: Region-based and query-based detectors · 2,060 citations (OpenAlex)

Check your idea against 108,967 papers

AI-extracted summary. Written by AI from the paper's title only (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 FPN with searched NAS-FPNRetinaNet with ResNet-50 · COCO test-dev, 640×640AP: 2.5 AP gainHelped
Use NAS-FPNLite instead of FPNLiteRetinaNet with MobileNetV2 · COCO test-dev, 320×320AP: 22.7 -> 24.2Helped
Use NAS-FPNLite instead of SSDLiteRetinaNet with MobileNetV2 · COCO test-dev, 320×320AP: 22.1 -> 24.2Helped
Increase stacked pyramid networks from 3 to 7 (ablation)NAS-FPNLite with MobileNetV2 · COCO test-dev, 320×320AP: 24.2 -> 25.7Helped
Add DropBlock (ablation)NAS-FPN with ResNet-50 (7 @ 384) · COCO test-dev, 1280×1280AP: 45.4 -> 46.6Helped
Add DropBlock (ablation)NAS-FPN with AmoebaNet (7 @ 384) · COCO test-dev, 1280×1280AP: 48.0 -> 48.3Helped
Stack repeated pyramid networks rather than use a single pyramid network (ablation)NAS-FPN · COCOAPHelped
Use deep supervision and early exit (ablation)NAS-FPN with ResNet-50 @ 384 · COCOAPNo clear effect

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