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

Deep Neural Networks (DNNs) that achieve human-level performance in generaltasks like object segmentation typically require supervised labels. Incontrast, humans are able to perform these tasks effortlessly withoutsupervision. To accomplish this, the human visual system makes use ofperceptual grouping. Understanding how perceptual grouping arises in anunsupervised manner is critical for improving both models of the visual system,and computer vision models. In this work, we propose a counterintuitiveapproach to unsupervised perceptual grouping and segmentation: that they arisebecause of neural noise, rather than in spite of it. We (1) mathematicallydemonstrate that under realistic assumptions, neural noise can be used toseparate objects from each other, and (2) show that adding noise in a DNNenables the network to segment images even though it was never trained on anysegmentation labels. Interestingly, we find that (3) segmenting objects usingnoise results in segmentation performance that aligns with the perceptualgrouping phenomena observed in humans. We introduce the Good Gestalt (GG)datasets -- six datasets designed to specifically test perceptual grouping, andshow that our DNN models reproduce many important phenomena in humanperception, such as illusory contours, closure, continuity, proximity, andocclusion. Finally, we (4) demonstrate the ecological plausibility of themethod by analyzing the sensitivity of the DNN to different magnitudes ofnoise. We find that some model variants consistently succeed with remarkablylow levels of neural noise ($\sigma<0.001$), and surprisingly, that segmentingthis way requires as few as a handful of samples. Together, our results suggesta novel unsupervised segmentation method requiring few assumptions, a newexplanation for the formation of perceptual grouping, and a potential benefitof neural noise in the visual system.

Latent Noise Segmentation: How Neural Noise Leads to the Emergence of Segmentation and Grouping