dAUTOMAP:分解AUTOMAP以实现可扩展性并提升性能
机器学习
2019-09-27 v2 图像与视频处理
机器学习
摘要
AUTOMAP是一种有前景的通用重建方法,然而其不可扩展,因而实用性受限。我们提出dAUTOMAP,一种分解AUTOMAP域变换的新方法,使模型呈线性规模扩展。我们展示dAUTOMAP以显著更少的参数优于AUTOMAP。
引用
@article{arxiv.1909.10995,
title = {dAUTOMAP: decomposing AUTOMAP to achieve scalability and enhance performance},
author = {Jo Schlemper and Ilkay Oksuz and James R. Clough and Jinming Duan and Andrew P. King and Julia A. Schnabel and Joseph V. Hajnal and Daniel Rueckert},
journal= {arXiv preprint arXiv:1909.10995},
year = {2019}
}
备注
Presented at ISMRM 27th Annual Meeting & Exhibition (Abstract #658)