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 Tools annual

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 Tools monthly

Global: Zendy Plus monthly

Presents the access of premium content as premium feature

Premium Content

Global: Zendy Plus annual

Global: Zendy Free Plan

Deep metric learning, which learns discriminative features to process imageclustering and retrieval tasks, has attracted extensive attention in recentyears. A number of deep metric learning methods, which ensure that similarexamples are mapped close to each other and dissimilar examples are mappedfarther apart, have been proposed to construct effective structures for lossfunctions and have shown promising results. In this paper, different from theapproaches on learning the loss structures, we propose a robust SNR distancemetric based on Signal-to-Noise Ratio (SNR) for measuring the similarity ofimage pairs for deep metric learning. By exploring the properties of our SNRdistance metric from the view of geometry space and statistical theory, weanalyze the properties of our metric and show that it can preserve the semanticsimilarity between image pairs, which well justify its suitability for deepmetric learning. Compared with Euclidean distance metric, our SNR distancemetric can further jointly reduce the intra-class distances and enlarge theinter-class distances for learned features. Leveraging our SNR distance metric,we propose Deep SNR-based Metric Learning (DSML) to generate discriminativefeature embeddings. By extensive experiments on three widely adoptedbenchmarks, including CARS196, CUB200-2011 and CIFAR10, our DSML has shown itssuperiority over other state-of-the-art methods. Additionally, we extend ourSNR distance metric to deep hashing learning, and conduct experiments on twobenchmarks, including CIFAR10 and NUS-WIDE, to demonstrate the effectivenessand generality of our SNR distance metric.

Signal-to-Noise Ratio: A Robust Distance Metric for Deep Metric Learning