Automation of leaf counting in maize and sorghum using deep learning. Issue 1 (27th August 2021)

• Record Type: Journal Article
• Authors: Miao, Chenyong; Guo, Alice; Thompson, Addie M.; Yang, Jinliang; Ge, Yufeng; Schnable, James C.
• Is Part Of: Plant phenome journal. Volume 4:Issue 1(2021)
• Page Start: n/a
• Page End: n/a
• Publication Date: 2021-08-27
• Subjects: Phenotype -- Periodicals; Plant genetics -- Periodicals; Periodicals; 581.35
• DOI: 10.1002/ppj2.20022
• Languages: English
• ISSNs: 2578-2703

Abstract

Abstract: Leaf number and leaf emergence rate are phenotypes of interest to plant breeders, plant geneticists, and crop modelers. Counting the extant leaves of an individual plant is straightforward even for an untrained individual, but manually tracking changes in leaf numbers for hundreds of individuals across multiple time points is logistically challenging. This study generated a dataset including over 150, 000 maize and sorghum images for leaf counting projects. A subset of 17, 783 images also includes annotations of the positions of individual leaf tips. With these annotated images, we evaluate two deep learning‐based approaches for automated leaf counting: the first based on counting‐by‐regression from whole image analysis and a second based on counting‐by‐detection. Both approaches can achieve root of mean square error (RMSE) smaller than one leaf, only moderately inferior to the RMSE between human annotators of between 0.57 and 0.73 leaves. The counting‐by‐regression approach based on convolutional neural networks (CNNs) exhibited lower accuracy and increased bias for plants with extreme leaf numbers which are underrepresented in this dataset. The counting‐by‐detection approach based on Faster R‐CNNs (region based convolutional neural networks) object detection models achieve near human performance for plants where all leaf tips are visible. The annotated image data and model performance metrics generated as part of this study provide large scale resources for the comparison and improvement of algorithms for leaf counting from image data in grain crops. Core Ideas: Automated leaf counting in maize was previously limited by annotated training data. Two deep learning approaches both achieve near human accuracy in counting leaves. There is an ongoing need to extend image datasets with phenotypically extreme plants.