Automated Image Registration: I. General Methods and Intrasubject, Intramodality Validation

TLDR

The study aims to describe and validate AIR 3.0, an automated image registration method that matches voxel intensities. The authors compared multiple cost functions, minimization methods, and sampling/smoothing/editing strategies, using internal consistency metrics for MRI and a brain phantom for PET to assess registration accuracy. All tested strategies achieved sub‑voxel accuracy, with MRI registration ranging from 75 to 150 µm, and sparse sampling cut processing time to minutes with only modest accuracy loss, demonstrating a robust, flexible tool adaptable to speed‑accuracy trade‑offs.

Abstract

Purpose We sought to describe and validate an automated image registration method(AIR 3.0) based on matching of voxel intensities. Method Different cost functions, different minimization methods, and various sampling, smoothing, and editing strategies were compared. Internal consistency measures were used to place limits on registration accuracy for MRI data, and absolute accuracy was measured using a brain phantom for PET data. Results All strategies were consistent with subvoxel accuracy for intrasubject, intramodality registration. Estimated accuracy of registration of structural MRI images was in the 75 to 150 μm range. Sparse data sampling strategies reduced registration times to minutes with only modest loss of accuracy. Conclusion The registration algorithm described is a robust and flexible tool that can be used to address a variety of image registration problems. Registration strategies can be tailored to meet different needs by optimizing tradeoffs between speed and accuracy.