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

MRI super-resolution (SR) and denoising tasks are fundamental challenges inthe field of deep learning, which have traditionally been treated as distincttasks with separate paired training data. In this paper, we propose aninnovative method that addresses both tasks simultaneously using a single deeplearning model, eliminating the need for explicitly paired noisy and cleanimages during training. Our proposed model is primarily trained for SR, butalso exhibits remarkable noise-cleaning capabilities in the super-resolvedimages. Instead of conventional approaches that introduce frequency-relatedoperations into the generative process, our novel approach involves the use ofa GAN model guided by a frequency-informed discriminator. To achieve this, weharness the power of the 3D Discrete Wavelet Transform (DWT) operation as afrequency constraint within the GAN framework for the SR task on magneticresonance imaging (MRI) data. Specifically, our contributions include: 1) a 3Dgenerator based on residual-in-residual connected blocks; 2) the integration ofthe 3D DWT with $1\times 1$ convolution into a DWT+conv unit within a 3D Unetfor the discriminator; 3) the use of the trained model for high-quality imageSR, accompanied by an intrinsic denoising process. We dub the model "DenoisingInduced Super-resolution GAN (DISGAN)" due to its dual effects of SR imagegeneration and simultaneous denoising. Departing from the traditional approachof training SR and denoising tasks as separate models, our proposed DISGAN istrained only on the SR task, but also achieves exceptional performance indenoising. The model is trained on 3D MRI data from dozens of subjects from theHuman Connectome Project (HCP) and further evaluated on previously unseen MRIdata from subjects with brain tumours and epilepsy to assess its denoising andSR performance.

DISGAN: Wavelet-informed Discriminator Guides GAN to MRI Super-resolution with Noise Cleaning