AI Paper Atlas

Research map / Architectures and efficiency / CNN design

Spatially-Adaptive Filter Units for Compact and Efficient Deep Neural Networks

Domen Tabernik, Matej Kristan, Aleš Leonardis

IJCV 2020 · Architectures and efficiency › CNN design · approach: Efficient CNN scaling and dynamic inference · 23 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
learn standard deviation per unit instead of fixing 0.5 (ablation)DAU-ConvNet · CIFAR10 (shallow 3-layer DAU-ConvNet)accuracy (%): 83.8 -> 84.25No clear effect
vary fixed standard deviation 0.3 to 0.8 (ablation)DAU-ConvNet · CIFAR10 (shallow 3-layer DAU-ConvNet)accuracy (%): 82.8 to 83.8No clear effect
prune units with relative threshold 0.01 to 0.02 (ablation)Large DAU-AlexNet · ILSVRC 2012 valtop-1 accuracy (%): 57.3 -> 57.1No clear effect
prune units with relative threshold 0.05 (ablation)Large DAU-AlexNet · ILSVRC 2012 valtop-1 accuracy (%): 57.3 -> 40.1Hurt
prune units with relative threshold 0.05 (ablation)Small DAU-AlexNet · ILSVRC 2012 valtop-1 accuracy (%): 56.4 -> 55.4Hurt
prune units with relative threshold 0.25 (ablation)Medium DAU-AlexNet · ILSVRC 2012 valtop-1 accuracy (%): 56.9 -> 1.9Hurt
replace conv filters by DAUs (Medium) in AlexNetAlexNet · ILSVRC 2012 valtop-1 accuracy (%): 56.99 -> 56.89 with 3.7 million -> 2.3 million paramsNo clear effect
replace 3x3 convs by two-unit DAUsResNet101 · ILSVRC 2012 valtop-1 accuracy (%): 75.39 -> 74.89Hurt
replace dilated convolution with DAUsAlexNet · PASCAL VOC 2011 valmIoU (%): 45.57 -> 47.22Improved
replace convolutions by DAUsResNet101 (OS16) · Cityscapes valmIoU (%): 68.6 -> 72.8Improved
add image-pooling features (ablation)ResNet101 · Cityscapes valmIoU (%): 68.6 -> 72.7Improved
replace ASPP with single DAU-6U layerDeepLabv3+ ResNet101 · Cityscapes valmIoU (%): 75.6 -> 75.5No clear effect

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