AUTHOR=Lee Da-Young , Na Dong-Yeop , Góngora-Canul Carlos , Baireddy Sriram , Lane Brenden , Cruz Andres P. , Fernández-Campos Mariela , Kleczewski Nathan M. , Telenko Darcy E. P. , Goodwin Stephen B. , Delp Edward J. , Cruz C. D. 

TITLE=Contour-Based Detection and Quantification of Tar Spot Stromata Using Red-Green-Blue (RGB) Imagery

JOURNAL=Frontiers in Plant Science

VOLUME=Volume 12 - 2021

YEAR=2021

URL=https://www.frontiersin.org/journals/plant-science/articles/10.3389/fpls.2021.675975

DOI=10.3389/fpls.2021.675975

ISSN=1664-462X

ABSTRACT=Quantifying tar spot intensity has traditionally been conducted through visual-based estimations of the proportion of leaf area covered by symptoms. However, this traditional method is costly in terms of time and labor as well as prone to rater subjectivity. An objective, accurate and high throughput method for quantifying stromata on tar-spot infected corn leaves is currently
unavailable. Here, we present contour-based detection of stromata for quantification of pathogen population using RGB images of tar spot-infected corn leaves. Image blocks (n = 1,130) generated by uniform partitioning the RGB images of leaves were assessed of the number and location of stromata using the newly developed stromata contour detection algorithm (SCDA; digital measurements) and two independent, experienced human raters using Image J (visual estimates). Stromata counts for each of the image blocks were then categorized into 5 classes and tested for the agreement of human raters and SCDA using Cohen’s weighted kappa coefficient. Nearly perfect agreement of stromata counts was observed for each of the human raters to SCDA as well as between the two human raters. Moreover, the SCDA was able to recognize ‘true stromata’ but to a lesser extent than human raters (average median recall=90.48 %, precision=89.69 %, and dice=88.31 %). Our results indicate the true potential utility of SCDA in quantifying stromata using RGB images, complementing the traditional human visual-based, disease severity estimations and serve as a foundation in building a high-throughput pipeline for tar spot of corn.