3D Motion Analysis in MRI Using a Multi-objective Evolutionary k-means Clustering

Many studies focused on gastric motility require the use of synthetic tracers to map the motion of content. Our study instead takes advantage of an unusual MRI acquisition protocol, combined with multi-objective optimised clustering to map the motion of food (peas, a natural ‘tracer’) in a human stomach. We chose NSGA-II to optimise the starting positions for a modified k-means to create optimum clusters. We compared our optimisation approach with a purely random approach that took an equal amount of processing time. Since we have no ground truth available, we have created alternative measures to evaluate our solutions: if the resulting pea velocities are within an expected range, and if each pea’s motion is correlated with neighbouring peas. We found that the optimised version has a significant improvement over the purely random search. Furthermore, we found many interesting food motion behaviours, such as correlated pea motion and more complex motion dynamics such as collision. Overall we found that the combined optimisation and clustering approach produced interesting findings relating to food dynamics in a human stomach.