AI Paper Atlas

Research map / Architectures and efficiency / CNN design

Analyzing Filters Toward Efficient ConvNet

Takumi Kobayashi

CVPR 2018 · Architectures and efficiency › CNN design · approach: Compact CNN architectures · 10 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 full convolution filters with orthonormal steerable bases from Analysis 1VGG-M · ILSVRC2014 validationtop-5 error rate (%): 12.96 -> 13.19Hurt
Use orthonormal steerable bases instead of random bases (ablation)VGG-M · ILSVRC2014 validationtop-5 error rate (%): 13.65 -> 13.19Improved
Require orthogonality in steerable filter bases (ablation)VGG-M · ILSVRC2014 validationtop-5 error rate (%): 13.64 -> 13.19Improved
Apply Analysis 1 convolution bases and dense-BoW max-poolingAlexNet · ILSVRC2014 validationtop-5 error rate (%): 16.58 -> 14.08Improved
Apply Analysis 1 convolution bases and dense-BoW max-poolingVGG-F · ILSVRC2014 validationtop-5 error rate (%): 16.36 -> 14.41Improved
Apply Analysis 1 convolution bases and dense-BoW average-poolingVGG-S · ILSVRC2014 validationtop-5 error rate (%): 13.19 -> 11.57Improved
Replace standard FC layer with dense-BoW average-pooling featuresVGG-vd-16 · ILSVRC2014 validationtop-5 error rate (%): 7.88 -> 7.10Improved
Replace standard FC layer with BoW featuresVGG-vd-16 · MIT67classification accuracy (%): 75.66 -> 79.12Improved
Replace standard FC layer with BoW featuresVGG-vd-16 · Bird200classification accuracy (%): 68.13 -> 77.60Improved
Replace conv5 and subsequent average-pooling with BoW (ablation)ResNet-50 · ILSVRC2014 validationtop-5 error rate (%): 6.35 -> 6.23Improved
Apply Analysis 1 3×3 bases at conv2–4 (ablation)ResNet-50 · ILSVRC2014 validationtop-5 error rate (%): 6.35 -> 6.42Hurt
Add Analysis 1 5×5 filters at conv4 (ablation)ResNet-50 · ILSVRC2014 validationtop-5 error rate (%): 6.31 -> 6.01Improved

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