TUdatalib : Content-Adaptive Downsampling in Convolutional Neural Networks

Zur Kurzanzeige

dc.contributor.author	Hesse, Robin
dc.contributor.author	Schaub-Meyer, Simone
dc.contributor.author	Roth, Stefan
dc.date.accessioned	2025-04-03T14:31:17Z
dc.date.available	2025-04-03T14:31:17Z
dc.date.issued	2023-06
dc.identifier.uri	https://tudatalib.ulb.tu-darmstadt.de/handle/tudatalib/4534
dc.description	Many convolutional neural networks (CNNs) rely on progressive downsampling of their feature maps to increase the network's receptive field and decrease computational cost. However, this comes at the price of losing granularity in the feature maps, limiting the ability to correctly understand images or recover fine detail in dense prediction tasks. To address this, common practice is to replace the last few downsampling operations in a CNN with dilated convolutions, allowing to retain the feature map resolution without reducing the receptive field, albeit increasing the computational cost. This allows to trade off predictive performance against cost, depending on the output feature resolution. By either regularly downsampling or not downsampling the entire feature map, existing work implicitly treats all regions of the input image and subsequent feature maps as equally important, which generally does not hold. We propose an adaptive downsampling scheme that generalizes the above idea by allowing to process informative regions at a higher resolution than less informative ones. In a variety of experiments, we demonstrate the versatility of our adaptive downsampling strategy and empirically show that it improves the cost-accuracy trade-off of various established CNNs.	de_DE
dc.language.iso	en	de_DE
dc.relation	IsDescribedBy;arXiv;2305.09504
dc.rights	Apache License 2.0
dc.rights.uri	https://www.apache.org/licenses/LICENSE-2.0
dc.subject	deep learning	de_DE
dc.subject	semantic segmentation	de_DE
dc.subject	keypoint estimation	de_DE
dc.subject	efficiency	de_DE
dc.subject.classification	4.43-05 Bild- und Sprachverarbeitung, Computergraphik und Visualisierung, Human Computer Interaction, Ubiquitous und Wearable Computing	de_DE
dc.subject.ddc	004
dc.title	Content-Adaptive Downsampling in Convolutional Neural Networks	de_DE
dc.type	Software	de_DE
dc.type	Model	de_DE
tud.project	EC/H2020 \| 866008 \| RED	de_DE
tud.project	HMWK \| 500/10.001-(00012) \| TAM - TP Roth	de_DE
tud.project	HMWK \| 519/03/06.001-(0010) \| WhiteBox - TP Roth	de_DE
tud.unit	TUDa