Embedding Gradient-based Optimization in Image Registration Networks
This work addresses image registration for medical imaging (e.g., cardiac and brain MR), offering a more efficient and robust solution, though it is incremental as it combines existing optimization and network approaches.
The paper tackled the problem of image registration by bridging traditional iterative optimization and deep learning methods, proposing GraDIRN, which embeds gradient-based optimization in a network's forward pass, achieving state-of-the-art performance with fewer parameters and good data efficiency and domain robustness.
Deep learning (DL) image registration methods amortize the costly pair-wise iterative optimization by training deep neural networks to predict the optimal transformation in one fast forward-pass. In this work, we bridge the gap between traditional iterative energy optimization-based registration and network-based registration, and propose Gradient Descent Network for Image Registration (GraDIRN). Our proposed approach trains a DL network that embeds unrolled multiresolution gradient-based energy optimization in its forward pass, which explicitly enforces image dissimilarity minimization in its update steps. Extensive evaluations were performed on registration tasks using 2D cardiac MR and 3D brain MR images. We demonstrate that our approach achieved state-of-the-art registration performance while using fewer learned parameters, with good data efficiency and domain robustness.