AI OnDemand: Segmenting at Scale from Napari

Poster presented as part of the Crick BioImage Analysis Symposium 2023.

Deep learning (DL) segmentation models provide the potential to greatly accelerate research, reducing manual annotation. Using these models can, however, require extensive work unique to each model. This is further complicated when working with HPC/cloud compute, which is often needed to run such models on real-world data. These models are generally not integrated into the usual work-flow/tooling, such as Napari, itself requiring additional work. Our solution, AI OnDemand, consists of a Napari plugin and Nextflow pipeline, seamlessly scaling compute as needed from within existing tools while providing easy access to various DL models.

Permission has been given by authors to upload to Crick Figshare. Copyright remains with the original authors.