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

TL;DR

This paper introduces a fast, global rigid multimodal 3D image alignment method using FFT-based normalized gradient fields, outperforming existing techniques in accuracy and speed, validated on brain imaging datasets.

Contribution

A novel FFT-based algorithm for computing NGF similarity in global multimodal 3D image alignment, enabling rapid and accurate registration.

Findings

Outperforms existing methods by a large margin.

Achieves alignment in approximately 40 seconds for 3.4 million voxels.

Validated on brain datasets with multiple modalities.

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 their initialization. We propose a global optimization method for rigid multimodal 3D image alignment, based on a novel efficient algorithm for computing similarity of normalized gradient fields (NGF) in the frequency domain. 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 all four reference methods by a large margin. The method is fast; a 3.4Mvoxel global rigid alignment 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 implementation is provided.