DeepPlate
Automated deep-learning pipeline for plate-based root and shoot phenotyping
Image
Deep learning
Python
Root architecture
Brachypodium
Arabidopsis
UniNe
Abstract
DeepPlate is an automated image-analysis pipeline for high-throughput phenotyping of plate-grown Brachypodium and Arabidopsis plants. Starting from high-resolution ScanStation images, it prepares a project workspace, crops user-defined regions (root, shoot, seed), and applies Ilastik-based deep-learning segmentation. Segmented images are then recomposed and used to extract quantitative traits such as projected area, width, depth, convex hull area and perimeter, enabling reproducible, scalable analysis of root and shoot architecture.
Background
On this page, I summarise the main performance metrics and example outputs of DeepPlate. I show how well the models segment different organs, how robust the pipeline is across experiments, and how the extracted traits relate to manual measurements.
Graphical overview
