Cross-Sim-NGF: FFT-Based Global Rigid Multimodal Alignment of Image Volumes using Normalized Gradient Fields

Johan Öfverstedt, Joakim Lindblad, Natasa Sladoje

Keywords: Image registration, global, exhaustive search, NGF, FFT, matching, GPU implementation

TL;DR: An efficient algorithm for computing similarities of NGF for all discrete displacements enabling fast global rigid alignment of 3D image volumes.

Abstract: Multimodal image alignment involves finding spatial correspondences between volumes varying in appearance and structure. Automated alignment methods are often based on local optimization that can be highly sensitive to initialization. We propose a novel efficient algorithm for computing similarity of normalized gradient fields (NGF) in the frequency domain, which we globally optimize to achieve rigid multimodal 3D image alignment. We validate the method experimentally on a dataset comprised of 20 brain volumes acquired in four modalities (T1w, Flair, CT, [18F] FDG PET), synthetically displaced with known transformations. The proposed method exhibits excellent performance on all six possible modality combinations and outperforms the four considered reference methods by a large margin. An important advantage of the method is its speed; global rigid alignment of 3.4\,Mvoxel volumes requires approximately 40 seconds of computation, and the proposed algorithm outperforms a direct algorithm for the same task by more than three orders of magnitude. Open-source code is provided.

Dataset Code:

The code is available at: http://github.com/MIDA-group/cross_sim_ngf.
The dataset we used is available upon request from Merida, I et al, related to the paper "CERMEP-IDB-MRXFDG: A database of 37 normal adult human brain [18F] FDG PET, T1 and FLAIR MRI, and CT images available for research" but we are not allowed to redistribute it.

Community Implementations: [ 1 code implementation](https://www.catalyzex.com/paper/cross-sim-ngf-fft-based-global-rigid/code)