AI Paper Atlas

Research map / Architectures and efficiency / CNN design

LIP: Local Importance-Based Pooling

Ziteng Gao, Limin Wang, Gangshan Wu

ICCV 2019 · Architectures and efficiency › CNN design · approach: Higher-order pooling · 141 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 LIP at the max pooling layer (ablation)LIP-ResNet-50 (Bottleneck-128) · ImageNet-1k valtop-1 accuracy: 77.87 -> 78.19Improved
add LIP at Res3 downsampling (ablation)LIP-ResNet-50 · ImageNet-1k valtop-1 accuracy: 77.78 -> 77.87Improved
add LIP at Res4 downsampling (ablation)LIP-ResNet-50 · ImageNet-1k valtop-1 accuracy: 76.92 -> 77.78Improved
add LIP at Res5 downsampling (ablation)LIP-ResNet-50 · ImageNet-1k valtop-1 accuracy: 76.40 -> 76.92Improved
add affine instance normalization to logit module (ablation)LIP-ResNet-50 · ImageNet-1k valtop-1 accuracy: 77.81 -> 78.19Improved
add amplified sigmoid to logit module (ablation)LIP-ResNet-50 · ImageNet-1k valtop-1 accuracy: 77.89 -> 78.19Improved
use Bottleneck-64 logit module instead of projection (ablation)LIP-ResNet-50 · ImageNet-1k valtop-1 accuracy: 77.49 -> 77.92Improved
widen logit module from Bottleneck-128 to Bottleneck-256 (ablation)LIP-ResNet-50 · ImageNet-1k valtop-1 accuracy: 78.19 -> 78.15No clear effect
replace average pooling with LIP Bottleneck-128ResNet-50 · ImageNet-1k valtop-1 accuracy: 76.96 -> 78.19Improved
replace strided convolution with average poolingResNet-50 · ImageNet-1k valtop-1 accuracy: 76.40 -> 76.96Improved
replace downsampling with LIPResNet-101 · ImageNet-1k valtop-1 accuracy: 77.98 -> 79.33Improved
replace downsampling with LIPDenseNet-BC-121 · ImageNet-1k valtop-1 accuracy: 75.62 -> 76.64Improved

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