AI Paper Atlas

Research map / Architectures and efficiency / CNN design

Deep Neural Networks with Box Convolutions

Egor Burkov, Victor Lempitsky

NeurIPS 2018 · Architectures and efficiency › CNN design · approach: Efficient CNN scaling and dynamic inference · 5 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 dilated/bottleneck blocks of ENet with box convolution blocks (BoxENet) (ablation)ENet · Cityscapes valIoU-class: 59.4 -> 64.6Improved
replace dilated/bottleneck blocks of ENet with box convolution blocks (BoxENet) (ablation)ENet · SUN RGB-D testmean IoU: 22.9% -> 24.5%Improved
replace blocks of ERFNet with box blocks (BoxERFNet) (ablation)ERFNet · Cityscapes valIoU-class: 68.8 -> 69.0No clear effect
replace blocks of ERFNet with box blocks (BoxERFNet) (ablation)ERFNet · SUN RGB-D testmean IoU: 25.3% -> 28.7%Improved
freeze box coordinates at initialization (ablation)BoxENet · Cityscapes valIoU-class: 64.6 -> 60.3Hurt
freeze box coordinates at initialization (ablation)BoxERFNet · Cityscapes valIoU-class: 69.0 -> 63.6Hurt
freeze box coordinates at initialization (ablation)BoxENet · SUN RGB-D testmean IoU: 24.5% -> 21.9%Hurt
remove residual blocks instead of replacing them with box blocks (ablation)ENet · Cityscapes valIoU-class: 54.1 vs 64.6 (BoxENet)Hurt
remove residual blocks instead of replacing them with box blocks (ablation)ERFNet · Cityscapes valIoU-class: 59.8 vs 69.0 (BoxERFNet)Hurt
replace all resolution-preserving blocks with box blocks (BoxOnlyENet) rather than every second (ablation)BoxENet · Cityscapes valIoU-class: 64.6 -> 61.0Hurt
use box convolutions in ERFNet family: computational cost (ablation)ERFNet · Resource costMultAdds (billions): 30.876 -> 16.042Improved
use box convolutions in ERFNet family: CPU inference (ablation)ERFNet · Resource costCPU time (ms): 1682 -> 822Improved

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