AI Paper Atlas

Research map / Detection and segmentation / CNN features

CARAFE: Content-Aware ReAssembly of FEatures

Jiaqi Wang, Kai Chen, Rui Xu, Ziwei Liu, Chen Change Loy, Dahua Lin

ICCV 2019 · Detection and segmentation › CNN features · approach: Joint detection and instance segmentation · 1,223 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 nearest upsampling in FPN with CARAFEFaster R-CNN R50 · COCO 2017 valbox AP: 36.5 -> 37.8Improved
add an extra 3x3 conv after nearest or bilinear upsamplingFaster R-CNN FPN · COCO 2017 valbox AP: 36.5 -> 36.6No clear effect
replace nearest upsampling in FPN with CARAFEFaster R-CNN R50 · COCO test-devbox AP: 36.9 -> 38.1Improved
replace nearest in FPN and deconv in mask head with CARAFEMask R-CNN R50 · COCO test-devmask AP: 34.6 -> 35.9Improved
replace deconvolution in the mask head with CARAFEMask R-CNN R50 · COCO 2017 valmask AP: 34.2 -> 34.7Improved
use CARAFE in both FPN and mask head instead of neitherMask R-CNN R50 · COCO 2017 valmask AP: 34.2 -> 35.7Improved
replace upsamplers in UperNet with CARAFEUperNet R50 · ADE20k valmIoU: 40.44 -> 42.23Improved
use CARAFE in all of PPM, FPN and FUSE rather than PPM only (ablation)UperNet R50 · ADE20k valmIoU: 40.85 -> 42.23Improved
replace upsampling with CARAFE in Global&Local inpaintingGlobal&Local · Places valPSNR (dB): 19.58 -> 20.71Improved
replace upsampling with CARAFE in Partial Conv inpaintingPartial Conv · Places valPSNR (dB): 20.78 -> 20.98Improved
compress channels to Cm=64 or remove the compressor (ablation)CARAFE · COCO 2017 valbox AP: 37.8 -> 37.8No clear effect
reduce compressed channels Cm from 64 to 16 (ablation)CARAFE · COCO 2017 valbox AP: 37.8 -> 37.6Hurt

Later papers that compared with this method

8 later accepted papers report CARAFE as a baseline in their results tables; 8 of them report a better number on most of the results they share with it.

Earlier methods this paper compares with

Named baselines from earlier accepted papers in this paper's results tables: better than 2 of 2 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