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

Intersection over Union (IoU) losses are surrogates that directly optimizethe Jaccard index. Leveraging IoU losses as part of the loss function havedemonstrated superior performance in semantic segmentation tasks compared tooptimizing pixel-wise losses such as the cross-entropy loss alone. However, weidentify a lack of flexibility in these losses to support vital trainingtechniques like label smoothing, knowledge distillation, and semi-supervisedlearning, mainly due to their inability to process soft labels. To addressthis, we introduce Jaccard Metric Losses (JMLs), which are identical to thesoft Jaccard loss in standard settings with hard labels but are fullycompatible with soft labels. We apply JMLs to three prominent use cases of softlabels: label smoothing, knowledge distillation and semi-supervised learning,and demonstrate their potential to enhance model accuracy and calibration. Ourexperiments show consistent improvements over the cross-entropy loss across 4semantic segmentation datasets (Cityscapes, PASCAL VOC, ADE20K, DeepGlobe Land)and 13 architectures, including classic CNNs and recent vision transformers.Remarkably, our straightforward approach significantly outperformsstate-of-the-art knowledge distillation and semi-supervised learning methods.The code is available at\href{https://github.com/zifuwanggg/JDTLosses}{https://github.com/zifuwanggg/JDTLosses}.

Jaccard Metric Losses: Optimizing the Jaccard Index with Soft Labels