Fast and Efficient Root Phenotyping via Pose Estimation

Plant Phenomics, Journal Year: 2024, Volume and Issue: 6

Published: Jan. 1, 2024

Image segmentation is commonly used to estimate the location and shape of plants their external structures. Segmentation masks are then localize landmarks interest compute other geometric features that correspond plant’s phenotype. Despite its prevalence, segmentation-based approaches laborious (requiring extensive annotation train) error-prone (derived sensitive instance mask integrity). Here, we present a segmentation-free approach leverages deep learning-based landmark detection grouping, also known as pose estimation. We use tool originally developed for animal motion capture called SLEAP (Social LEAP Estimates Animal Poses) automate distinct morphological on plant roots. Using gel cylinder imaging system across multiple species, show our can reliably efficiently recover root topology at high accuracy, few annotated samples, faster speed than approaches. In order make this landmark-based representation phenotyping, Python library ( sleap-roots ) trait extraction directly comparable existing analysis software. pose-derived traits highly accurate be common downstream tasks including genotype classification unsupervised mapping. Altogether, work establishes validity advantages estimation-based phenotyping. To facilitate adoption easy-to-use encourage further development, , all training data, models, code available at: https://github.com/talmolab/sleap-roots https://osf.io/k7j9g/ .

Language: Английский

ClearDepth: a simple, robust, and low‐cost method to assess root depth in soil

The Plant Journal, Journal Year: 2024, Volume and Issue: unknown

Published: Dec. 8, 2024

SUMMARY Root depth is a major determinant of plant performance during drought and key trait for strategies to improve soil carbon sequestration mitigate climate change. While the model Arabidopsis thaliana offers numerous advantages studies root system architecture depth, its small fragile roots severely limit use methods techniques currently available such in soils. To overcome this, we have developed ClearDepth, conceptually simple, non‐destructive, sensitive, low‐cost method estimate relatively pots that are amenable mid‐ large‐scale studies. In our method, develops naturally inside soil, without considerable space constraints. The ClearDepth parameter wall shallowness (WRS) quantifies by measuring reach transparent walls clear pots. We show WRS robust sensitive distinguishes deep systems from shallower ones while also capturing smaller differences caused influence an environmental factor. addition, leveraged study relation between lateral angles measured non‐soil soil. found genotypes characterized steep growth media produce deeper Finally, can be used crop species like rice.

Language: Английский