Deep learning based geometric registration for medical images: How accurate can we get without visual features?
This work addresses the challenge of geometric registration in medical imaging, offering a novel approach that could improve accuracy for tasks like lung structure alignment, though it is incremental in shifting focus from visual to geometric features.
The authors tackled the problem of medical image registration by developing a deep learning framework that relies solely on geometric features and optimization, rather than visual features, and achieved highly accurate 3D point cloud registration, strongly outperforming dense encoder-decoder networks and other point set registration methods on complex key-point graphs of inner lung structures.
As in other areas of medical image analysis, e.g. semantic segmentation, deep learning is currently driving the development of new approaches for image registration. Multi-scale encoder-decoder network architectures achieve state-of-the-art accuracy on tasks such as intra-patient alignment of abdominal CT or brain MRI registration, especially when additional supervision, such as anatomical labels, is available. The success of these methods relies to a large extent on the outstanding ability of deep CNNs to extract descriptive visual features from the input images. In contrast to conventional methods, the explicit inclusion of geometric information plays only a minor role, if at all. In this work we take a look at an exactly opposite approach by investigating a deep learning framework for registration based solely on geometric features and optimisation. We combine graph convolutions with loopy belief message passing to enable highly accurate 3D point cloud registration. Our experimental validation is conducted on complex key-point graphs of inner lung structures, strongly outperforming dense encoder-decoder networks and other point set registration methods. Our code is publicly available at https://github.com/multimodallearning/deep-geo-reg.