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 Plus monthly

Global: Zendy Tools annual

Global: Zendy Tools monthly

Global: Zendy Free Plan

Recent findings show that deep convolutional neural networks (DCNNs) do notgeneralize well under partial occlusion. Inspired by the success ofcompositional models at classifying partially occluded objects, we propose tointegrate compositional models and DCNNs into a unified deep model with innaterobustness to partial occlusion. We term this architecture CompositionalConvolutional Neural Network. In particular, we propose to replace the fullyconnected classification head of a DCNN with a differentiable compositionalmodel. The generative nature of the compositional model enables it to localizeoccluders and subsequently focus on the non-occluded parts of the object. Weconduct classification experiments on artificially occluded images as well asreal images of partially occluded objects from the MS-COCO dataset. The resultsshow that DCNNs do not classify occluded objects robustly, even when trainedwith data that is strongly augmented with partial occlusions. Our proposedmodel outperforms standard DCNNs by a large margin at classifying partiallyoccluded objects, even when it has not been exposed to occluded objects duringtraining. Additional experiments demonstrate that CompositionalNets can alsolocalize the occluders accurately, despite being trained with class labelsonly. The code used in this work is publicly available.

Compositional Convolutional Neural Networks: A Deep Architecture with Innate Robustness to Partial Occlusion