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

Perceiving the world in terms of objects and tracking them through time is acrucial prerequisite for reasoning and scene understanding. Recently, severalmethods have been proposed for unsupervised learning of object-centricrepresentations. However, since these models were evaluated on differentdownstream tasks, it remains unclear how they compare in terms of basicperceptual abilities such as detection, figure-ground segmentation and trackingof objects. To close this gap, we design a benchmark with four data sets ofvarying complexity and seven additional test sets featuring challengingtracking scenarios relevant for natural videos. Using this benchmark, wecompare the perceptual abilities of four object-centric approaches: ViMON, avideo-extension of MONet, based on recurrent spatial attention, OP3, whichexploits clustering via spatial mixture models, as well as TBA and SCALOR,which use explicit factorization via spatial transformers. Our results suggestthat the architectures with unconstrained latent representations learn morepowerful representations in terms of object detection, segmentation andtracking than the spatial transformer based architectures. We also observe thatnone of the methods are able to gracefully handle the most challenging trackingscenarios despite their synthetic nature, suggesting that our benchmark mayprovide fruitful guidance towards learning more robust object-centric videorepresentations.

Benchmarking Unsupervised Object Representations for Video Sequences