Multi-task MR Imaging with Iterative Teacher Forcing and Re-weighted Deep Learning

Abstract

Noises, artifacts, and loss of information caused by the magnetic resonance (MR) reconstruction may compromise the final performance of the downstream applications. In this paper, we develop a re-weighted multi-task deep learning method to learn prior knowledge from the existing big dataset and then utilize them to assist simultaneous MR reconstruction and segmentation from the under-sampled k-space data. The multi-task deep learning framework is equipped with two network sub-modules, which are integrated and trained by our designed iterative teacher forcing scheme (ITFS) under the dynamic re-weighted loss constraint (DRLC). The ITFS is designed to avoid er-

A large, curated, open-source stroke neuroimaging dataset to improve lesion segmentation algorithms

“…The ATLAS v2.0 dataset was developed using similar protocols and methods as the ATLAS v1.2 dataset, which has been successfully utilized to develop numerous lesion segmentation methods for the last several years [12][13][14][15][16][17][18][19][20][21][22][23][24][25][26][27][28] . For ATLAS v2.0, detailed manual quality control for image quality occurred during the initial lesion segmentation, and all segmentations were examined for quality by two additional researchers.…”