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

Image classifiers are information-discarding machines, by design. Yet, howthese models discard information remains mysterious. We hypothesize that oneway for image classifiers to reach high accuracy is to first zoom to the mostdiscriminative region in the image and then extract features from there topredict image labels, discarding the rest of the image. Studying six popularnetworks ranging from AlexNet to CLIP, we find that proper framing of the inputimage can lead to the correct classification of 98.91% of ImageNet images.Furthermore, we uncover positional biases in various datasets, especially astrong center bias in two popular datasets: ImageNet-A and ObjectNet. Finally,leveraging our insights into the potential of zooming, we propose a test-timeaugmentation (TTA) technique that improves classification accuracy by forcingmodels to explicitly perform zoom-in operations before making predictions. Ourmethod is more interpretable, accurate, and faster than MEMO, astate-of-the-art (SOTA) TTA method. We introduce ImageNet-Hard, a new benchmarkthat challenges SOTA classifiers including large vision-language models evenwhen optimal zooming is allowed.

ImageNet-Hard: The Hardest Images Remaining from a Study of the Power of Zoom and Spatial Biases in Image Classification