Spotiflow is a deep learning method for subpixel-accurate spot detection in fluorescence microscopy images. It formulates spot detection as a multiscale heatmap and stereographic flow regression problem, achieving state-of-the-art performance on diverse datasets while being robust to different noise conditions.
Key features: - Subpixel-accurate 2D and 3D spot detection - Optimized for spatial transcriptomics workflows - Robust to low signal-to-noise ratio images - Available as Python library and napari plugin
Find the GitHub repo here