'empanada: what’s cool, what’s new and what’s next'
with Dr Kedar Narayan
About this Webinar
Description:
empanada is a napari plugin for panoptic segmentation workflows on 2D and 3D cellular EM images (https://empanada.readthedocs.io/en/latest/); it is designed to be used in resource- and expertise- limited settings. It allows facile model training, fine-tuning, inference and proof-reading, and comes with pre-trained generalist instance segmentation models MitoNet, NucleoNet and DropNet. Here I discuss some of the background work and resources associated with creating empanada, and some examples of how it has enabled new insights in cell biology. I will include a simple show-and-tell tutorial on how to run empanada, including some features included in the latest release.
Speaker biography:
Kedar Narayan is a senior scientist and group leader at the Center for Cancer Research Volume Electron Microscopy (CVEM) at Frederick National Laboratory and National Cancer Institute, USA. Kedar has a Ph.D. in immunology, with an emphasis on biophysics and imaging, and a background in chemistry, pathology and software engineering. His group has developed and applied FIB-SEM and other volume EM technologies to questions in cancer and cell biology. Specific areas of Kedar’s research focus are correlative imaging, vEM tool development and deep learning/AI; his lab released “empanada”, a popular napari plugin for automated segmentation of organelles from EM images. He has co-authored more than sixty papers and given invited talks internationally in this space. His community work includes co-organizing conferences on volume EM and “large data”, leadership on data working groups, and creating common metadata standards for the field. For his scientific innovation and leadership in the volume EM field, Kedar was awarded the Alan Agar Award by the Royal Microscopy Society in 2025. As a leading member of the volumeEM community, Kedar is committed to the growth and democratization of the field.
Intended audience:
Learning outcomes:
The talk will be broadly accessible and is aimed at scientists of all levels in and adjacent to the fields of cell biology, imaging and computational/AI-based analysis
Level:
No prior knowledge is needed.
Resources:
A basic understanding of the concepts of AI based segmentation and analysis of volume EM images
An ability to apply empanada and other napari plugins to my own EM data
empanada documentation: https://empanada.readthedocs.io/en/latest/