AI Paper Atlas

Research map / Architectures and efficiency / CNN design

DFT-based Transformation Invariant Pooling Layer for Visual Classification

Jongbin Ryu, Ming-Hsuan Yang, Jongwoo Lim

ECCV 2018 · Architectures and efficiency › CNN design · approach: Filter design and trained-filter analysis · 46 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 average pooling/direct FC with DFT magnitude poolingVGG-VD16 · ImageNet (from scratch)top-1 error: 29.09 -> 27.28Improved
replace average pooling with DFT magnitude poolingResNet-50 · ImageNet (from scratch)top-1 error: 25.15 -> 24.37Improved
replace pooling with DFT magnitude poolingAlexNet · ImageNet (from scratch)top-1 error: 41.12 -> 40.23Improved
add mid-layer fusion (DFT+) to DFT poolingVGG-VD16 · ImageNet (from scratch)top-1 error: 27.28 -> 27.07Improved
add mid-layer ensemble with average pooling (average+)ResNet-50 · ImageNet (from scratch)top-1 error: 25.15 -> 25.13No clear effect
use learnable conv initialised as 2D-DFT instead of explicit DFT (ablation)AlexNet · ImageNet (from scratch)top-1 error: 40.23 -> 40.64Hurt
use random-init learnable conv instead of explicit DFT (ablation)AlexNet · ImageNet (from scratch)top-1 error: 40.23 -> 40.71Hurt
replace average pooling with DFT magnitude poolingAlexNet · Caltech 101 (fine-tuned)accuracy: 88.1 -> 87.4Hurt
replace average pooling with DFT magnitude poolingResNet-50 · CUB (fine-tuned)accuracy: 77.5 -> 81.0Improved
replace average pooling with DFT magnitude poolingResNet-50 · MIT Indoor (fine-tuned)accuracy: 73.0 -> 74.8Improved
reduce DFT pooling size to N=2 (ablation)ResNet-50 · CUB (fine-tuned)accuracy: 77.5 -> 76.2Hurt
increase DFT pooling size from N=2 to full (ablation)ResNet-50 · CUB (fine-tuned)accuracy: 76.2 -> 81.0Improved

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