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

The ability to estimate the perceptual error between images is an importantproblem in computer vision with many applications. Although it has been studiedextensively, however, no method currently exists that can robustly predictvisual differences like humans. Some previous approaches used hand-codedmodels, but they fail to model the complexity of the human visual system.Others used machine learning to train models on human-labeled datasets, butcreating large, high-quality datasets is difficult because people are unable toassign consistent error labels to distorted images. In this paper, we present anew learning-based method that is the first to predict perceptual image errorlike human observers. Since it is much easier for people to compare two givenimages and identify the one more similar to a reference than to assign qualityscores to each, we propose a new, large-scale dataset labeled with theprobability that humans will prefer one image over another. We then train adeep-learning model using a novel, pairwise-learning framework to predict thepreference of one distorted image over the other. Our key observation is thatour trained network can then be used separately with only one distorted imageand a reference to predict its perceptual error, without ever being trained onexplicit human perceptual-error labels. The perceptual error estimated by ournew metric, PieAPP, is well-correlated with human opinion. Furthermore, itsignificantly outperforms existing algorithms, beating the state-of-the-art byalmost 3x on our test set in terms of binary error rate, while alsogeneralizing to new kinds of distortions, unlike previous learning-basedmethods.

PieAPP: Perceptual Image-Error Assessment through Pairwise Preference