Using a Structural Root System Model to Evaluate and Improve the Accuracy of Root Image Analysis Pipelines

Abstract

Root system analysis is a complex task, often performed with fully automated image analysis pipelines. However, the outcome is rarely verified by ground-truth data, which might lead to underestimated biases. We have used a root model, ArchiSimple, to create a large and diverse library of ground-truth root system images (10,000). For each image, three levels of noise were created. This library was used to evaluate the accuracy and usefulness of several image descriptors classically used in root image analysis softwares. Our analysis highlighted that the accuracy of the different traits is strongly dependent on the quality of the images and the type, size, and complexity of the root systems analyzed. Our study also demonstrated that machine learning algorithms can be trained on a synthetic library to improve the estimation of several root system traits. Overall, our analysis is a call to caution when using automatic root image analysis tools. If a thorough calibration is not performed on the dataset of interest, unexpected errors might arise, especially for large and complex root images. To facilitate such calibration, both the image library and the different codes used in the study have been made available to the community.

Existing and Potential Statistical and Computational Approaches for the Analysis of 3D CT Images of Plant Roots

“…An approach to addressing this issue is to use simulated data to improve accuracy in root image analysis pipelines [24]. Improving structural root system simulation and how to combine simulated data with real data in root system analysis are open research questions for further study.…”

“…The study of such interactions is enabled by non-destructive phenotyping technologies that can examine the phenotypic impacts of such interactions in soil during the stages of growth [22]. Such impacts are captured by phenotypic traits, i.e., qualitative and quantitative measures that describe the appearance of plants and their roots [23,24].…”

Latent Space Phenotyping: Automatic Image-Based Phenotyping for Treatment Studies

“…Synthetic images of rosettes [32] and roots [20] have been used previously to train models for phenotyping tasks. Here we used synthetically generated image data as it allows us to introduce specific variance in the imagery based on a simulated casual SNP, and then investigate the method's ability to recover that variance on the other end by running a GWAS on the simulated population.…”

Combining semi-automated image analysis techniques with machine learning algorithms to accelerate large scale genetic studies

“…The improvement of root architectural traits will thus be crucial in delivering the yield improvement required to ensure future global food security [1,2]. Unfortunately, root systems are difficult to analyse and quantify: they are intrinsically complex due to their highly branched tree structure [3], and their growth in an opaque medium (soil) makes them difficult to observe. For many years, root researchers have used specific experimental setups to observe and quantify root system architecture.…”

“…However, such user interaction is time consuming, which can strongly hinder the application of such approaches to large datasets (such as those required for quantitative genetic studies). Fully automated software tools are faster, but the extracted descriptors are prone to unexpected errors and the quantified traits are usually less informative [3]. This has led to image analysis being described as a new "bottleneck" in plant phenotyping [11].…”