Self-attention-based non-linear basis transformations for compact latent space modelling of dynamic optical fibre transmission matrices
Researchers have developed a novel approach using self-attention layers to model dynamic optical fibre transmission matrices. This method aims to improve image reconstruction for next-generation medical endoscopes by addressing the inherent scrambling of light within multimode optical fibres. The self-attention mechanism transforms the coordinate representations of these matrices into a basis that allows for compact, low-dimensional representations, significantly enhancing sparsity and enabling reconstruction with low error. AI
IMPACT This research could lead to more accurate and detailed internal imaging for medical applications by improving how light signals are processed.