Interpolation Artefacts in Mutual Information-Based Image Registration

By Josien P. W. Pluim, J. B. Antoine Maintz, Max A. Viergever. Computer Vision and Image Understanding, 2000

308 citations recorded by the source corpus.

Subfields: Computer Vision and Pattern Recognition

View Paper: https://www.semanticscholar.org/paper/2c5450a75979c5eec152c8abb36d83f68c9a520b

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Computer Vision and Pattern Recognition

Citations in the corpus

An approach to multimodal biomedical image registration utilizing particle swarm optimization (2004, methodology): “Partial volume interpolation [36], in which the intensity histogram is updated using weighted intensities of neighboring voxels, was used in this study, as this method was found to reduce transients and artifacts in the similarity metric function [36].”
3D Deep Learning on Medical Images: A Review (2020, background): “The image alignment not only makes it easier to manually analyze images and locate lesions or other abnormalities, but also makes it easier to train a 3D CNN on these images [48–50].”
elastix: A Toolbox for Intensity-Based Medical Image Registration (2010, background): “It has often been studied in the context of interpolation artefacts [25].”
A global optimisation method for robust affine registration of brain images (2001, background): “These results the precise effects of interpolation is an active research show that the proposed method is more reliable at finding area (Hajnal et al., 1995; Pluim et al., 2000; Thacker et al., the global minimum than the other packages.”
Intensity Gradient Based Registration and Fusion of Multi-modal Images (2006, background): “However, mutual information has a number of well-known drawbacks; cf. e.g., [17], [16], [19].”
Motion Correction Algorithms May Create Spurious Brain Activations in the Absence of Subject Motion (2001, background): “While MI was used for historical reasons during our work, this may not be the best choice for motion correction, first because of computational time considerations, second because recent results have shown that MI is prone to local maxima problems (Roche et al., 1998; Studholme et al., 1999; Pluim ”
What is the best similarity measure for motion correction in fMRI time series? (2002, background): “result from the interpolation artifact observed by Pluim in [34] and have then no impact on real data.”
A survey of medical image registration on graphics hardware (2011, background): “The exact interpolation used is important, as it can lead to artifacts and influences the similarity measure [3,4].”
Image registration by maximization of combined mutual information and gradient information (2000, background): “For such translations, interpolation is not applied, resulting in a higher joint entropy because of the presence of noise and a decreased mutual information value (see [15] for a more detailed explanation).”
Mutual-information-based registration of medical images: a survey (2003, background): “In [132], the different patterns created by linear and partial volume interpolation are extensively studied.”