Unsupervised Multi-Modality Registration Network based on Spatially Encoded Gradient Information
This work addresses multi-modality medical image registration, a key problem for clinical applications like organ or tumor analysis, though it appears incremental as it builds on existing neural network approaches.
The authors tackled the challenge of registering multi-modality medical images by proposing MMRegNet, which uses spatially encoded gradient information for unsupervised training and achieved promising performance on left ventricle cardiac registration tasks from the MM-WHS 2017 dataset, with additional validation on a liver dataset from CHAOS 2019.
Multi-modality medical images can provide relevant or complementary information for a target (organ, tumor or tissue). Registering multi-modality images to a common space can fuse these comprehensive information, and bring convenience for clinical application. Recently, neural networks have been widely investigated to boost registration methods. However, it is still challenging to develop a multi-modality registration network due to the lack of robust criteria for network training. In this work, we propose a multi-modality registration network (MMRegNet), which can perform registration between multi-modality images. Meanwhile, we present spatially encoded gradient information to train MMRegNet in an unsupervised manner. The proposed network was evaluated on MM-WHS 2017. Results show that MMRegNet can achieve promising performance for left ventricle cardiac registration tasks. Meanwhile, to demonstrate the versatility of MMRegNet, we further evaluate the method with a liver dataset from CHAOS 2019. Source code will be released publicly\footnote{https://github.com/NanYoMy/mmregnet} once the manuscript is accepted.