基于动态卷积的多模态脑MR图像生成  

Multi-modality Brain MR Images Synthesis Based on Dynamic Convolution

作　　者：孙君顶[1] 杨鸿章 闫艺丹 毋小省[1] 唐朝生[1] SUN Jun-Ding;YANG Hong-Zhang;YAN Yi-Dan;WU Xiao-Sheng;TANG Chao-Sheng(School of Computer Science and Technology,Henan Polytechnic University,Jiaozuo 454000,China)

机构地区：[1]河南理工大学计算机科学与技术学院,焦作454000

出　　处：《计算机系统应用》2022年第8期305-313,共9页Computer Systems & Applications

基　　金：河南省科技厅科技攻关项目(212102310084);河南省高等学校重点科研项目(22A520027)。

摘　　要：近年来,通过自动生成方法获取多模态MR图像得到了广泛研究,但仍难以通过一种模态直接生成其他各类模态的图像.针对该问题,本文提出了动态生成对抗网络.新模型通过将生成对抗网络与动态卷积相结合,同时加入任务标签这一条件,实现从一种MR模态同时生成其他3种MR模态.同时为了提高图像生成质量,进一步提出了多尺度判别策略,通过融合多个尺度来提升判别效果.基于BRATS19数据集进行生成验证,实验结果表明,新方法不但可以同时生成多种模态的数据,而且提高了生成图像的质量.In recent years,obtaining multi-modality magnetic resonance(MR)images with automatic generation methods has been widely studied.However,it is still difficult to generate images of all the other modalities by one given modality.To solve this problem,this study proposes a dynamic generative adversarial network(DyGAN)model.By combining the generative adversarial network and dynamic convolution and introducing a task label,the new model can simultaneously generate other three MR modalities from one modality.In addition,a multi-scale discrimination strategy is further proposed to improve the quality of image generation by fusing multiple scales.Image generation is verified on the BRATS19 dataset.The experimental results show that the new method can not only simultaneously generate multimodality images but also improve the quality of the generated images.

关 键 词：图像生成 多模态 动态卷积 动态生成对抗网络 多尺度判别器 深度学习 生成对抗网络 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术] R445.2[自动化与计算机技术—计算机科学与技术] R739.41[医药卫生—影像医学与核医学]