AI Paper Atlas

Research map / Detection and segmentation / CNN features

FADE: A Task-Agnostic Upsampling Operator for Encoder–Decoder Architectures

Hao Lü, Wenze Liu, Hongtao Fu, Zhiguo Cao

IJCV 2024 · Detection and segmentation › CNN features · approach: Local features and context in CNNs · 17 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 encoder features to decoder-only kernel generation (ablation)CARAFE (b2) · ADE20K SegFormer-B1mIoU: 42.82 -> 43.27Improved
add decoder features to encoder-only kernel generation (ablation)b1 · ADE20K SegFormer-B1mIoU: 42.75 -> 43.27Improved
use semi-shift convolution instead of naive kernel generation (ablation)FADE · ADE20K SegFormer-B1mIoU: 43.27 -> 43.33Improved
use semi-shift convolution instead of naive kernel generation (ablation)FADE · Composition-1KSAD: 32.41 -> 31.78Improved
add gating instead of skipping encoder features (ablation)FADE · ADE20K SegFormer-B1mIoU: 43.22 -> 44.41Improved
add gating mechanism (ablation)FADE b4 · ADE20K SegFormer-B1mIoU: 43.33 -> 44.41Improved
add gate to nearest-neighbor upsampling (ablation)SegNet NN · Weizmann HorsemIoU: 89.15 -> 95.26Improved
add gate to CARAFE (ablation)SegNet CARAFE · Weizmann HorsemIoU: 89.29 -> 95.25Improved
use iterative x2 upsampling instead of direct (ablation)FADE · ADE20K SegFormer-B1mIoU: 42.89 -> 44.41Improved
add FADE to strong SegFormer-B5 backboneSegFormer-B5 · ADE20KmIoU: 51.00 -> 50.90No clear effect
add FADE to SegFormer-B5 backbone, boundary metricSegFormer-B5 · ADE20KbIoU: 37.81 -> 38.91Improved
add gating to FADE for object detection (ablation)FADE · COCO Faster R-CNN R50AP: 38.5 -> 37.8Hurt

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