Wang Yun, Chang Wanru, Huang Chongfei, Kong Dexing
School of Mathematical Sciences, Zhejiang University, Hangzhou, China.
College of Mathematics and Computer Science, Zhejiang A&F University, Hangzhou, China.
J Xray Sci Technol. 2024;32(6):1385-1398. doi: 10.3233/XST-240159.
Deformable image registration (DIR) plays an important part in many clinical tasks, and deep learning has made significant progress in DIR over the past few years.
To propose a fast multiscale unsupervised deformable image registration (referred to as FMIRNet) method for monomodal image registration.
We designed a multiscale fusion module to estimate the large displacement field by combining and refining the deformation fields of three scales. The spatial attention mechanism was employed in our fusion module to weight the displacement field pixel by pixel. Except mean square error (MSE), we additionally added structural similarity (ssim) measure during the training phase to enhance the structural consistency between the deformed images and the fixed images.
Our registration method was evaluated on EchoNet, CHAOS and SLIVER, and had indeed performance improvement in terms of SSIM, NCC and NMI scores. Furthermore, we integrated the FMIRNet into the segmentation network (FCN, UNet) to boost the segmentation task on a dataset with few manual annotations in our joint leaning frameworks. The experimental results indicated that the joint segmentation methods had performance improvement in terms of Dice, HD and ASSD scores.
Our proposed FMIRNet is effective for large deformation estimation, and its registration capability is generalizable and robust in joint registration and segmentation frameworks to generate reliable labels for training segmentation tasks.
可变形图像配准（DIR）在许多临床任务中发挥着重要作用，并且在过去几年中深度学习在DIR方面取得了显著进展。
提出一种用于单模态图像配准的快速多尺度无监督可变形图像配准（称为FMIRNet）方法。
我们设计了一个多尺度融合模块，通过组合和细化三个尺度的变形场来估计大位移场。我们的融合模块采用空间注意力机制对位移场进行逐像素加权。除了均方误差（MSE），我们在训练阶段还额外添加了结构相似性（ssim）度量，以增强变形图像与固定图像之间的结构一致性。
我们的配准方法在EchoNet、CHAOS和SLIVER上进行了评估，在SSIM、NCC和NMI分数方面确实有性能提升。此外，我们将FMIRNet集成到分割网络（FCN、UNet）中，以在我们的联合学习框架中提高在人工标注较少的数据集上的分割任务性能。实验结果表明，联合分割方法在Dice、HD和ASSD分数方面有性能提升。
我们提出的FMIRNet对于大变形估计是有效的，并且其配准能力在联合配准和分割框架中具有通用性和鲁棒性，能够为训练分割任务生成可靠的标签。
