基于自注意力机制生成对抗网络的超分辨率磁共振图像重建

Super-resolution reconstruction of MR image with self-attention based generate adversarial network algorithm

高分辨率的磁共振图像可以提供细粒度的解剖信息,但是获取数据需要较长的扫描时间.本文提出了一种基于自注意力机制生成对抗网络的超分辨率磁共振图像重构方法(SA-SR-GAN),利用生成对抗网络从低分辨率磁共振图像生成高分辨率磁共振图像,将自注意力机制集成到超分辨率生成对抗网络框架中,用于计算输入特征的权重参数,同时引入了谱归一化处理,使判别器网络训练过程更加稳定.本文使用40组3D磁共振图像(每组图像包含256个切片)训练网络,并用10组图像进行测试.实验结果表明,所提出的超分辨率自注意力生成对抗网络方法生成的超分辨率的磁共振图像的PSNR和SSIM值高于同类比较方法.

Super-resolution(SR)MRI images can provide fine-grained anatomical information,however it takes a long time to acquire data.In order to accelerate the acquisition of MR images while maintaining high-quality images,extensive research has been performed on image reconstruction through the deep learning method.In this study,a reconstruction framework by using self-attention based super-resolution generative adversarial networks(SA-SR-GAN)is proposed to generate super resolution MR image from low resolution MR image.Moreover,the self-attention mechanism is integrated into super-resolution generative adversarial networks(SR-GAN)framework,which is used to calculate the weight parameters of the input features.At the same time,spectral normalization is added to make the discriminator network training process more stable.The network was trained with 403D images(each 3D image contains 256 slices)and tested with 10 images.The experimental results show that the PSNR and SSIM values of the super-resolution magnetic resonance image generated by the proposed SA-SR-GAN method are higher than the state-of-the-art reconstruction methods.