Latent flow improves speed and accuracy in medical image segmentation
Latent-to-Latent Flow for Volumetric Stochastic Segmentation
Computer Vision and Pattern RecognitionArtificial IntelligenceMachine Learning
Summary
Medical images often have uncertain boundaries because different doctors may outline structures differently. This uncertainty can make planning treatments harder. The researchers developed a new method that uses a shortcut by working with compressed versions of the images and labels, which makes the process much faster. Their technique keeps accuracy high while being up to 14 times quicker than older methods. It helps safely capture differences in how organs or treatment areas are seen on scans.
medical image segmentationinter-observer variabilityvolumetric dataflow matchinglatent representationstochastic segmentationradiotherapy planningmultiple organ segmentationgenerative modeling
Authors
Omar Todd, Sooha Kim, Raghav Mehta, Katherine Mackay, David Bernstein, Alexandra Taylor, Fabio De Sousa Ribeiro, Ben Glocker
Abstract
Uncertainty arising from inter-observer variability in medical image segmentation plays an important role in developing treatment plans. Research in this area is inhibited by the lack of multiple annotations for large-scale medical datasets, especially for volumetric data, which suffers from additional scaling and computational complexity challenges. Flow matching has emerged as a powerful framework for generative modelling and has also been demonstrated to maintain strong performance when working with latent representations of images. In this work, we introduce a latent-to-latent flow technique for stochastic segmentation of medical volumes via encoded representations of both the image and label space. We evaluate our method on two challenging applications covering delineation uncertainty for radiotherapy planning and multiple organ structure segmentation, improving efficiency up to 14x compared with full resolution models while maintaining clinically relevant performance.