AI Paper Atlas

Research map / Detection and segmentation / CNN features

PT-CapsNet: A Novel Prediction-Tuning Capsule Network Suitable for Deeper Architectures

Chenbin Pan, Senem Velipasalar

ICCV 2021 · Detection and segmentation › CNN features · approach: Convolution-transformer backbones · 18 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 capsule-wise prediction then feature-wise tuning instead of the reverse order (ablation)PT-CapsNet small model · MNISTtesting error: 0.75 -> 0.63Improved
add capsule-wise prediction then feature-wise tuning instead of the reverse order (ablation)PT-CapsNet small model · Fashion-MNISTtesting error: 8.29 -> 7.77Improved
add generated logits (extra FC-PT-Caps layer) instead of L2-norm logits (ablation)PT-CapsNet small model · MNISTtesting error: 88.65 -> 0.63Improved
use ReLU instead of Squash with generated logits (ablation)PT-CapsNet small model · MNISTtesting error: 0.74 -> 0.63Improved
replace conv and FC layers with LC-PT-Caps and FC-PT-CapsResNet-110 · CIFAR-10top-1 accuracy: 92.69 -> 93.59Improved
replace conv and FC layers with LC-PT-Caps and FC-PT-CapsWRN-28-10 · CIFAR-10top-1 accuracy: 95.95 -> 95.98No clear effect
replace conv and FC layers with LC-PT-Caps and FC-PT-CapsDenseNet-100 · CIFAR-10top-1 accuracy: 95.33 -> 95.71Improved
replace conv and FC layers with PT-Caps layersResNet-110 · CIFAR-100top-1 accuracy: 70.32 -> 71.39Improved
replace conv and FC layers with PT-Caps layersDenseNet-100 · CIFAR-100top-1 accuracy: 77.12 -> 78.36Improved
replace conv and FC layers with PT-Caps layersResNet-110 · FashionMNISTtop-1 accuracy: 94.36 -> 95.28Improved
replace conv layers with PT-Caps layersU-Net · ISIC2018mIoU: 65.57 -> 66.86Improved
replace conv layers with PT-Caps layersDeepLabv3+ · ISIC2018mIoU: 82.85 -> 83.12Improved

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