Problem
Research questions and friction points this paper is trying to address.
shift-invariance
stretch-invariance
non-negative matrix factorization
emission tomography
dynamic neuroimaging
Shift- and stretch-invariant non-negative matrix factorization with an application to brain tissue delineation in emission tomography data
🤖 AI Summary
Dynamic neuroimaging data are often compromised by time delays, scale mismatches, and temporal stretching induced by diffusion effects, which severely limit the performance of conventional linear modeling and decomposition approaches. This work proposes a translation- and stretch-invariant non-negative matrix factorization framework that, for the first time, jointly estimates both integer and non-integer temporal shifts and stretching factors in the frequency domain: temporal translations are addressed through phase adjustments, while temporal stretching is modeled via zero-padding or truncation. The method demonstrates robust efficacy on both synthetic data and brain emission tomography datasets, significantly improving the accuracy of cerebral tissue structure delineation and establishing a novel paradigm for dynamic neuroimaging analysis.
📝 Abstract
Dynamic neuroimaging data, such as emission tomography measurements of radiotracer transport in blood or cerebrospinal fluid, often exhibit diffusion-like properties. These introduce distance-dependent temporal delays, scale-differences, and stretching effects that limit the effectiveness of conventional linear modeling and decomposition methods. To address this, we present the shift- and stretch-invariant non-negative matrix factorization framework. Our approach estimates both integer and non-integer temporal shifts as well as temporal stretching, all implemented in the frequency domain, where shifts correspond to phase modifications, and where stretching is handled via zero-padding or truncation. The model is implemented in PyTorch (https://github.com/anders-s-olsen/shiftstretchNMF). We demonstrate on synthetic data and brain emission tomography data that the model is able to account for stretching to provide more detailed characterization of brain tissue structure.
Innovation
Methods, ideas, or system contributions that make the work stand out.
shift-invariant
stretch-invariant
non-negative matrix factorization
frequency domain
dynamic neuroimaging
🔎 Similar Papers
No similar papers found.
