Learned shape representations for segmentation of cells and organelles in light and electron microscopy images

Session

Stream 2: Software and Smart Microscopy

Authors

Martin Weigert (1)

Affiliations

1. EPFL

Keywords

Image Analysis, Deep Learning, Segmentation

Abstract text

Automatic detection and segmentation of cells, nuclei, and organelles in light and electron microscopy images is often the first step in many biological analysis pipelines.
Typically employed machine-learning based approaches however can struggle in cases of many crowded objects where separation of individual instances is challenging, a situation that is quite common in biological imaging experiments.
In my talk, I will highlight our ongoing efforts in developing user friendly deep learning based detection and segmentation methods that address these issues by using learned star-convex shape representations. I will demonstrate the merits of this approach on examples ranging from nuclei detection in fluorescence microscopy, tissue classification in histopathology to the comprehensive organelle reconstruction of whole cells in 3D electron microscopy.

References

Schmidt, U., Weigert, M., Broaddus, C., & Myers, G. (2018, September). Cell detection with star-convex polygons. In International Conference on Medical Image Computing and Computer-Assisted Intervention (pp. 265-273). Springer, Cham.

Weigert, M., Schmidt, U., Haase, R., Sugawara, K., & Myers, G. (2020). Star-convex polyhedra for 3d object detection and segmentation in microscopy. In Proceedings of the IEEE/CVF Winter Conference on Applications of Computer Vision (pp. 3666-3673).

Müller, Andreas, et al. "3D FIB-SEM reconstruction of microtubule–organelle interaction in whole primary mouse β cells." Journal of Cell Biology 220.2 (2021).