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 potential of deep neural networks in skin lesion classification hasalready been demonstrated to be on-par if not superior to the dermatologistsdiagnosis. However, the performance of these models usually deteriorates whenthe test data differs significantly from the training data (i.e. domain shift).This concerning limitation for models intended to be used in real-world skinlesion classification tasks poses a risk to patients. For example, differentimage acquisition systems or previously unseen anatomical sites on the patientcan suffice to cause such domain shifts. Mitigating the negative effect of suchshifts is therefore crucial, but developing effective methods to address domainshift has proven to be challenging. In this study, we carry out an in-depthanalysis of eight different unsupervised domain adaptation methods to analyzetheir effectiveness in improving generalization for dermoscopic datasets. Toensure robustness of our findings, we test each method on a total of tendistinct datasets, thereby covering a variety of possible domain shifts. Inaddition, we investigated which factors in the domain shifted datasets have animpact on the effectiveness of domain adaptation methods. Our findings showthat all of the eight domain adaptation methods result in improved AUPRC forthe majority of analyzed datasets. Altogether, these results indicate thatunsupervised domain adaptations generally lead to performance improvements forthe binary melanoma-nevus classification task regardless of the nature of thedomain shift. However, small or heavily imbalanced datasets lead to a reducedconformity of the results due to the influence of these factors on the methodsperformance.

Mitigating the Influence of Domain Shift in Skin Lesion Classification:  A Benchmark Study of Unsupervised Domain Adaptation Methods on Dermoscopic  Images