Primitive Representation Learning for Unsupervised Dynamic Contrast Enhanced MRI Reconstruction
2026-08-18 • Computer Vision and Pattern Recognition
Computer Vision and Pattern RecognitionMachine Learning
AI summaryⓘ
The authors developed a new way to improve MRI images that show how organs change over time after a contrast dye is used. Their method breaks down the image data into parts: the basic body structure, the way the contrast dye moves, and any leftover motion. This helps create clearer images even when the scan is very fast and not many data points are collected. They found their approach works as well as existing methods for seeing how blood moves through organs like the aorta and kidneys.
dynamic contrast-enhanced MRIimage reconstructionGaussian primitivesGabor primitivesspatiotemporal reconstructiontemporal basis functionsundersamplingcontrast enhancementmodular designacceleration rates
Authors
Veronika Spieker, Wenqi Huang, Cemre Ariyurek, Liam Timms, Daniel Rueckert, Onur Afacan, Julia A. Schnabel, Sila Kurugol
Abstract
Reliable quantitative analysis of dynamic contrast-enhanced MRI requires high-quality spatiotemporal reconstructions at high undersampling rates. Scan-specific reconstructions using Gaussian and Gabor primitives have shown promising results without the need for large training datasets, but have not addressed the additional dimension of dynamic contrast. We propose a multi-dimensional, primitive based framework for dynamic contrast-enhanced MRI reconstruction that disentangles the underlying anatomy, the dynamic contrast enhancement, and residual motion into separate temporal basis functions, thereby enabling a geometrical interpretation of the representation. We show that this architecture achieves performance competitive with conventional reconstruction methods, both in reconstruction quality and in the accuracy of extracted aorta and kidney enhancement curves. The modular tier design extends naturally to additional dynamic factors and higher acceleration rates. Code available at https://github.com/compai-lab/ 2026-GaborDCE-spieker.