Spotiflow: accurate and efficient spot detection for fluorescence microscopy with deep stereographic flow regression

Albert Dominguez Mantes, Antonio Herrera, Irina Khven, Anjalie Schlaeppi, Eftychia Kyriacou, Georgios Tsissios, Evangelia Skoufa, Luca Santangeli, Ekaterina Buglakova, Emine Berna Durmus, Suliana Manley, Anna Kreshuk, Detlev Arendt, Can Aztekin, Joachim Lingner, Gioele La Manno, Martin Weigert

Abstract

The identification of spot-like structures in large and noisy microscopy images is an important task in many life science techniques. Imaging-based spatial transcriptomics (iST) methods rely critically on the accurate detection of millions of transcripts in low signal-to-noise ratio (SNR) images. Here we present Spotiflow, a deep learning method for subpixel-accurate spot detection that formulates spot detection as a multiscale heatmap and stereographic flow regression problem. Spotiflow supports 2D and 3D images, generalizes across different imaging conditions and is more time and memory efficient than existing methods. We demonstrate that Spotiflow achieves state-of-the-art performance on diverse fluorescence microscopy datasets while being robust to different noise conditions.