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

Multiwavelength photoacoustic images encode information about a tissue'soptical absorption distribution. This can be used to estimate its blood oxygensaturation distribution (sO2), an important physiological indicator of tissuehealth and pathology. However the wavelength dependence of the light fluencedistribution complicates the recovery of accurate estimates, in particular,preventing the use of a straightforward spectroscopic inversion. Deep learningapproaches have been shown effective at producing accurate estimates of sO2from simulated data. Though, the translation of generic supervised learningapproaches to real tissues is prevented by the lack of real `paired' trainingdata (multiwavelength PA images of in vivo tissues with their corresponding sO2distributions). Here, we discuss i) why networks trained on images simulatedusing conventional means are unlikely to generalise their performance on realtissues, and ii) the prospects of using two generative adversarial networkbased strategies to improve the generalisability of sO2-estimating networkstrained on synthetic data: a) CycleGAN-driven unsupervised domain adaptation ofconventionally simulated images, and b) the generation of paired training datausing AmbientGANs.

Enhancing synthetic training data for quantitative photoacoustic tomography with generative deep learning