(leer)

Navigation and service


A Deep Learning Framework for Transforming Image Reconstruction Into Pixel Classification

5 February 2020

Kamlesh Pawar, Zhaolin Chen, N. Jon Shah, Gary. F. Egan

Since the inception of MRI, the main conundrum faced by researchers is how to get the best possible image in as little time as possible.

In this collaborative study led by our colleagues at Monash University, Melbourne, Australia, a method using a deep learning framework is presented that transforms the task of MR image reconstruction from randomly undersampled k-space data into pixel classification, thus improving speed and accuracy.

It was found that the reconstructed images using the DL classification model outperformed the state-of-the-art compressed sensing and DL regression models with a similar number of parameters assessed using quantitative measures. Furthermore, the experiments conducted in this research show that the proposed deep learning method is robust to noise and is able to reconstruct high-quality images in low SNR scenarios where conventional CS reconstructions and DL regression networks perform poorly.

The proposed approach is generic in nature and demonstrates that many of the advances made in deep learning classification problems can be integrated into MR image reconstruction tasks and has the potential to leverage great advances in this area of research.

Original publication:

A Deep Learning Framework for Transforming Image Reconstruction Into Pixel Classification


Servicemeu

Homepage