Truzman et al., (in prep)
Manual Lesion Tracing: The established gold standard for lesion segmentation involves expert clinicians manually delineating lesions on MRI scans, a process that can be time-consuming and susceptible to inter-rater variability.
Semi-Automated Lesion Segmentation Tools: While researchers havProjects e developed (semi-)automated tools for identifying lesions in stroke MRI scans, their performance often decreases when applied to data acquired from different locations or environments.
Automated AI Lesion Segmentation: We have developed a fully automated artificial intelligence tool for segmenting lesions in both acute and chronic stroke, achieving performance comparable to expert clinicians with improved speed and objectivity.
Open Science Approach: Our developed models and code are openly accessible, accompanied by comprehensive documentation to facilitate their use by the wider research community.