Related Experiment Video
Updated: Jul 9, 2025
11:37
RGB and Spectral Root Imaging for Plant Phenotyping and Physiological Research: Experimental Setup and Imaging Protocols
Published on: August 8, 2017
16.2K
Fast and efficient root phenotyping via pose estimation
Elizabeth M Berrigan1, Lin Wang1, Hannah Carrillo1
1Salk Institute for Biological Studies, La Jolla, CA 92037 United States of America.
Biorxiv : the Preprint Server for Biology
|December 4, 2023
Summary
This study introduces a new deep learning method for plant root phenotyping, using pose estimation to accurately map root structures faster and with less data than traditional segmentation methods.
Area of Science:
- Plant biology
- Computer vision
- Bioinformatics
Background:
- Traditional plant phenotyping relies on image segmentation, which is labor-intensive and prone to errors.
- Segmentation-based methods require extensive data annotation for training and are sensitive to mask imperfections.
Approach:
- A novel, segmentation-free approach using deep learning-based pose estimation (Social LEAP Estimates Animal Poses - SLEAP) for plant root analysis.
- SLEAP automates the detection of morphological landmarks on plant roots, overcoming limitations of segmentation.
Key Points:
- The approach achieves high accuracy in recovering root system topology with fewer annotated samples and increased speed.
- A Python library, sleap-roots, enables trait extraction directly from landmarks, comparable to segmentation-based software.
- Landmark-derived traits are accurate for downstream tasks like genotype classification and trait mapping.
Conclusions:
- Pose estimation offers a more efficient and reliable alternative to segmentation for plant phenotyping.
- The developed tool (sleap-roots) and resources are publicly available to promote adoption and further research.
