ChronoRoot 2.0: An Open AI-Powered Platform for 2D Temporal Plant Phenotyping

TL;DR

ChronoRoot 2.0 presents an open-source, AI-powered platform for 2D temporal plant phenotyping that combines low-cost hardware with a self-configuring nnUNet for multi-class segmentation of six plant structures. The work introduces dual graphical interfaces for detailed RSA analysis and high-throughput screening, integrates FPCA for temporal pattern discovery, and demonstrates robust performance across Arabidopsis and tomato, including under diverse growth conditions. Key advancements include improved segmentation accuracy, automated seed tracking, and a scalable framework for single-plant and multi-plant analyses, all supported by RSML export and open data resources. The platform enables researchers without computational expertise to perform sophisticated temporal RSA analyses, supporting scalable throughput and cross-species comparisons with potential applications in plant breeding and functional genomics.

Abstract

Plant developmental plasticity, particularly in root system architecture, is fundamental to understanding adaptability and agricultural sustainability. ChronoRoot 2.0 builds upon established low-cost hardware while significantly enhancing software capabilities and usability. The system employs nnUNet architecture for multi-class segmentation, demonstrating significant accuracy improvements while simultaneously tracking six distinct plant structures encompassing root, shoot, and seed components: main root, lateral roots, seed, hypocotyl, leaves, and petiole. This architecture enables easy retraining and incorporation of additional training data without requiring machine learning expertise. The platform introduces dual specialized graphical interfaces: a Standard Interface for detailed architectural analysis with novel gravitropic response parameters, and a Screening Interface enabling high-throughput analysis of multiple plants through automated tracking. Functional Principal Component Analysis integration enables discovery of novel phenotypic parameters through temporal pattern comparison. We demonstrate multi-species analysis, with Arabidopsis thaliana and Solanum lycopersicum, both morphologically distinct plant species. Three use cases in Arabidopsis thaliana and validation with tomato seedlings demonstrate enhanced capabilities: circadian growth pattern characterization, gravitropic response analysis in transgenic plants, and high-throughput etiolation screening across multiple genotypes.ChronoRoot 2.0 maintains the low-cost, modular hardware advantages of its predecessor while dramatically improving accessibility through intuitive graphical interfaces and expanded analytical capabilities. The open-source platform makes sophisticated temporal plant phenotyping more accessible to researchers without computational expertise.

Figures (10)

• Figure 1: ChronoRoot 2.0: An integrated platform for temporal plant phenotyping. (A) The hardware module combines affordable components for automated imaging in controlled environments. (B) Infrared images are captured continuously, enabling consistent monitoring during both day and night cycles. (C) A multi-class segmentation model based on nnUNet automatically identifies and tracks six plant structures: main root, lateral roots, seed, hypocotyl, leaves, and petiole. The system routes data through an interface selection step, offering two specialized analysis workflows: (D) the Standard Interface for detailed architectural analysis of individual plants, and (E) the Screening Interface for high-throughput experiments involving multiple individuals.
• Figure 2: Angular measurements in root system architecture. Illustration of base-tip angle ($\theta_{bt}$) and emergence angle ($\theta_{e}$) calculations on a Arabidopsis thaliana plant showing how these complementary metrics quantify different aspects of lateral root orientation.
• Figure 3: Use Case 1 - Arabidopsis thaliana Root system architecture dynamics under different light conditions. Comparison of long-day (16h/8h, blue, n=23) versus continuous light (24h, orange, n=21) shows divergent growth patterns. A - All basic Architectural RSA parameters, B - FPCA analysis of Main Root Length, significative differences found in PC2 (p-value<0.001), C - Fourier transform of Total Root Growth Rate (significative differences found at both 24h and 12h periods, p-value<0.05), D - FPCA analysis of Lateral Root Length, significative differences found in PC2 (p-value<0.05). Error bands: standard error.
• Figure 4: Use Case 2 - Altered root architecture in Arabidopsis thaliana NF-YA10miRres plants. A - Qualitative analysis of convex hull showing root system coverage differences between genotypes. B - Quantitative analysis of convex hull metrics (area and aspect ratio) between NF-YA10miRres plants (1 in orange, n=26), 2 in green, n=28) and Col0 controls (blue, n=16). C - Quantitative analysis of average emergence angle and base-tip angle for the first lateral root, demonstrating consistently wider angles in NF-YA10miRres plants compared to Col0. Error bars: standard error.
• Figure 5: Use Case 3 - High-throughput analysis of Arabidopsis thaliana seedling etiolation. A - Representative infrared images showing temporal progression of etiolated seedling development. B - Hypocotyl, root and total area measurements with their corresponding growth rates (G1, blue, n=189. G2, orange, n=211; G3, green, n=180). C - Germination curves showed T50 at approximately 13 hs after light stimulation with no significant differences between the analyzed genotypes. D - Comparison of Manual (blue) and automatic (orange) hypocotyl length determinations at 36, 48 and 60 hours after light stimulation (no significant difference observed). Error bars: standard error.