English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Global: Zendy Plus annual

Presents the access of premium content as premium feature

Premium Content

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Global: Zendy Tools annual

Global: Zendy Free Plan

Global: Zendy Tools monthly

Global: Zendy Plus monthly

We study the first-order scattering transform as a candidate for reducing thesignal processed by a convolutional neural network (CNN). We show theoreticaland empirical evidence that in the case of natural images and sufficientlysmall translation invariance, this transform preserves most of the signalinformation needed for classification while substantially reducing the spatialresolution and total signal size. We demonstrate that cascading a CNN with thisrepresentation performs on par with ImageNet classification models, commonlyused in downstream tasks, such as the ResNet-50. We subsequently apply ourtrained hybrid ImageNet model as a base model on a detection system, which hastypically larger image inputs. On Pascal VOC and COCO detection tasks wedemonstrate improvements in the inference speed and training memory consumptioncompared to models trained directly on the input image.

Compressing the Input for CNNs with the First-Order Scattering Transform