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

Ultrasound is the second most used modality in medical imaging. It is costeffective, hazardless, portable and implemented routinely in numerous clinicalprocedures. Nonetheless, image quality is characterized by granulatedappearance, poor SNR and speckle noise. Specific for malignant tumors, themargins are blurred and indistinct. Thus, there is a great need for improvingultrasound image quality. We hypothesize that this can be achieved, usingneural networks, by translation into a more realistic display which mimics ananatomical cut through the tissue. In order to achieve this goal, thepreferable approach would be to use a set of paired images. However, this ispractically impossible in our case. Therefore, Cycle Generative AdversarialNetwork (CycleGAN) was used, in order to learn each domain propertiesseparately and enforce cross domain cycle consistency. The two datasets whichwere used for training the model were "Breast Ultrasound Images" (BUSI) and aset of optic images of poultry breast tissue samples acquired at our lab. Thegenerated pseudo anatomical images provide improved visual discrimination ofthe lesions with clearer border definition and pronounced contrast. In order toevaluate the preservation of the anatomical features, the lesions in theultrasonic images and the generated pseudo anatomical images were bothautomatically segmented and compared. This comparison yielded median dice scoreof 0.91 for the benign tumors and 0.70 for the malignant ones. The medianlesion center error was 0.58% and 3.27% for the benign and malignanciesrespectively and the median area error index was 0.40% and 4.34% for the benignand malignancies respectively. In conclusion, these generated pseudo anatomicalimages, which are presented in a more intuitive way, enhance tissue anatomy andcan potentially simplify the diagnosis and improve the clinical outcome.

Image translation of Ultrasound to Pseudo Anatomical Display by CycleGAN