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

Crafting effective deep learning models for medical image analysis is acomplex task, particularly in cases where the medical image dataset lackssignificant inter-class variation. This challenge is further aggravated whenemploying such datasets to generate synthetic images using generativeadversarial networks (GANs), as the output of GANs heavily relies on the inputdata. In this research, we propose a novel filtering algorithm called CosineSimilarity-based Image Filtering (CosSIF). We leverage CosSIF to develop twodistinct filtering methods: Filtering Before GAN Training (FBGT) and FilteringAfter GAN Training (FAGT). FBGT involves the removal of real images thatexhibit similarities to images of other classes before utilizing them as thetraining dataset for a GAN. On the other hand, FAGT focuses on eliminatingsynthetic images with less discriminative features compared to real images usedfor training the GAN. Experimental results reveal that employing either theFAGT or FBGT method with modern transformer and convolutional-based networksleads to substantial performance gains in various evaluation metrics. FAGTimplementation on the ISIC-2016 dataset surpasses the baseline method in termsof sensitivity by 1.59% and AUC by 1.88%. Furthermore, for the HAM10000dataset, applying FABT outperforms the baseline approach in terms of recall by13.75%, and with the sole implementation of FAGT, achieves a maximum accuracyof 94.44%.

CosSIF: Cosine similarity-based image filtering to overcome low inter-class variation in synthetic medical image datasets