This readme.txt file was generated on 2025-05-16 by Julian Cooper. Recommended citation for the data: Cooper, Julian; Du, Chuan; Beaver, Zach; Zheng, Ming; Page, Rae; Wodarek, Joseph; Matny, Oadi; Szinyei, Tomas; Quiñones, Alejandra; Anderson, James; Smith, Kevin; Yang, Ce; Steffenson, Brian; Hirsch, Cory. (2025). An RGB deep-neural network approach for high-throughput phenotyping of Fusarium head blight in wheat. Retrieved from the Data Repository for the University of Minnesota (DRUM), https://doi.org/10.13020/8JWK-XX41. ------------------- GENERAL INFORMATION ------------------- 1. Title of Dataset: An RGB deep-neural network approach for high-throughput phenotyping of Fusarium head blight in wheat 2. Author Information Author Contact: Cory Hirsch (cdhirsch@umn.edu) Name: Julian Cooper Institution: University of Minnesota Email: coop0409@umn.edu ORCID: 0000-0001-8390-3915 Name: Chuan Du Institution: Mineral Earth Science LLC Email: chuandu@x.team ORCID: NA Name: Zach Beaver Institution: Mineral Earth Science LLC Email: zbeaver@x.team ORCID: NA Name: Ming Zheng Institution: Mineral Earth Science LLC Email: zhengming@google.com ORCID: NA Name: Rae Page Institution: University of Minnesota Email: pagex277@umn.edu ORCID: 0009-0009-5516-7354 Name: Joseph Wodarek Institution: University of Minnesota Email: wodar006@crk.umn.edu ORCID: NA Name: Oadi Matny Institution: University of Minnesota Email: onmatny@umn.edu ORCID: 0000-0002-8447-2886 Name: Tomas Szinyei Institution: University of Minnesota Email: tszinyei@umn.edu ORCID: NA Name: Alejandra Quiñones Institution: University of Minnesota Email: quino070@umn.edu ORCID: 0000-0002-0603-491X Name: James Anderson Institution: University of Minnesota Email: ander319@umn.edu ORCID: 0000-0003-4655-6517 Name: Kevin Smith Institution: University of Minnesota Email: smith376@umn.edu ORCID: 0000-0001-8253-3326 Name: Ce Yang Institution: University of Minnesota Email: ceyang@umn.edu ORCID: 0000-0002-1079-118X Name: Brian Steffenson Institution: University of Minnesota Email: bsteffen@umn.edu ORCID: 0000-0001-7961-5363 Name: Cory Hirsch Institution: University of Minnesota Email: cdhirsch@umn.edu ORCID: 0000-0002-3409-758X 3. Date published or finalized for release: 2025-04-08 4. Date of data collection: 2022-06-01 to 2022-08-01 5. Geographic location of data collection (where was data collected?): NA 6. Information about funding sources that supported the collection of the data: This material is based upon work supported by the U.S. Department of Agriculture, under Project numbers: 59-0206-1-198 and 59-0206-2-127. This is a cooperative project with the U.S. Wheat & Barley Scab Initiative. Any opinions, findings, conclusions, or recommendations expressed in this publication are those of the authors and do not necessarily reflect the view of the U.S. Department of Agriculture. This project was also supported by the University of Minnesota Experimental Station USDA-NIFA Hatch project MIN-22-086. Julian Cooper was supported by Bayer Crop Sciences as a Bayer Crop Sciences Fellow. 7. Overview of the data (abstract): Fusarium head blight (FHB) in wheat is an economically important disease which can cause yield losses exceeding 50%. Breeding for host resistance remains the most effective disease control method; however, time, labor, and human subjectivity during disease scoring limits selection advancements. In this study we describe an innovative, high-throughput phenotyping rover for capturing in-field RGB images and a deep neural network pipeline for wheat spike detection and FHB disease quantification. The image analysis pipeline detects wheat spikes from images collected by the phenotyping rover under variable field conditions, segments those spikes and the amount of diseased tissue in the spikes, and quantifies disease severity as the region of intersection between the spike and disease masks. To validate disease inferences, individual spike and plot aggregate FHB estimates from the pipeline were compared with visual disease scores from the field and on images. The precision and throughput of the pipeline surpassed traditional field rating methods. Aggregate plot disease levels as estimated by the pipeline correlated highly with field and manually annotated image disease scores; however disease assessments on individual spikes were influenced by field location. The pipeline was able to quantify FHB from images taken with different camera orientations than the original training data, which demonstrates strong generalizability. This innovative pipeline represents a breakthrough in FHB phenotyping, offering precise and efficient assessment of FHB on both individual spikes and plot aggregates. The pipeline is robust across different environments and the potential to standardize disease evaluation methods across the research groups make it a valuable tool for studying and managing this economically significant fungal disease. -------------------------- SHARING/ACCESS INFORMATION -------------------------- 1. Licenses/restrictions placed on the data: CC0 1.0 Universal (https://creativecommons.org/publicdomain/zero/1.0/) 2. Links to publications that cite or use the data: Julian Cooper, Chuan Du, Zach Beaver, Ming Zheng, Rae Page, Joseph R. Wodarek, Oadi Matny, Tamas Szinyei, Alejandra Quiñones, James A. Anderson, Kevin P. Smith, Ce Yang, Brian J. Steffenson, Cory D. Hirsch. (2023). An RGB based deep neural network for high fidelity Fusarium head blight phenotyping in wheat. bioRxiv 2023.09.20.558703; doi: https://doi.org/10.1101/2023.09.20.558703. 3. Was data derived from another source? No If yes, list source(s): 4. Terms of Use: Data Repository for the U of Minnesota (DRUM) By using these files, users agree to the Terms of Use. https://conservancy.umn.edu/pages/policies/#drum-terms-of-use -------------------------- METHODOLOGICAL INFORMATION -------------------------- 1. Description of methods used for collection/generation of data: Images were collected using the terrestrial phenotyping rover developed by Mineral Earth Science, LLC. 2. Methods for processing the data: Data was processed and disease inferences were generated using a deep learning image analysis pipeline in Python. 3. Instrument- or software-specific information needed to interpret the data: NA 4. Standards and calibration information, if appropriate: NA 5. Environmental/experimental conditions: Images of diseases were collected from two locations in Saint Paul and Crookston, Minnesota from 2021 and 2022. Experiments were artificially inoculated and irrigated during initial disease infection. 6. Describe any quality-assurance procedures performed on the data: Disease inferences were validated against raters scoring disease in the field and manually from images. 7. People involved with sample collection, processing, analysis and/or submission: Julian Cooper, Alejandra Quiñones, Oadi Matny, Tamas Szinyei, and Rae Page were involved in data collection. Julian Cooper, Chuan Du, Zach Beaver, and Ming Zheng were involved in data processing and analysis. Julian Cooper was involved in data submission. --------------------- DATA & FILE OVERVIEW --------------------- 1_fhb_disease_data: Fusarium head blight field and image data frames and analysis - FHB_disease_human_field: For each row/plot at each rating date/location, 0-100% plot average disease scores from five raters, A-E. - FHB_disease_pipeline_image: For each row/plot at each rating date/location, individual spike disease percentage (fhb_percentage) and gradability classification (gradability_prob, <0.50 = gradable) for all detected wheat heads. - FHB_disease_human_pipeline_200image: For the 200 spike images used to measure inter-rater reliability, individual spike disease percentage from each rater, A-E, and from the FHB image analysis pipeline (Pipeline). - FHB_disease_human_pipeline_10Kimage: For all spikes determined to be gradable by raters during manual image annotation of the 10K images used to compare human and pipeline disease annotations on a spike and plot level, individual spike disease percentage from the FHB image analysis pipeline (pipeline_image_perc_FHB) and from manual image annotation by a human rater (human_image_perc_FHB) as well as the rater ID. 2_image_analysis_pipeline_workflow: Image analysis pipeline for analyzing images 3_wheat_spike_detection_model: Contains grain_head_and_other_detection model for wheat spike detection to be loaded into image analysis pipeline 4_wheat_spike_gradability_model: Contains umn_full_growth_cycle_unet_128x128 model for wheat spike gradability classification to be loaded into image analysis pipeline 5_wheat_spike_segmentation_model: Contains wheat_head_fhb_seg model for wheat spike segmentation to be loaded into image analysis pipeline 6_disease_segmentation_model: Contains wheat_spike_fhb_disease_non_gradable_classification_input_is_BGR model for disease segmentation to be loaded into image analysis pipeline 7_n200_inter_rater_reliability_images: Five raters and pipeline annotating disease on same 200 spikes to compare inter-rater reliability Files are labelled as {date}_{location}_{ID}_bin{bin}_rater{rater}.png Where ID is unique image ID Bin in the disease score per the N10K image annotation subset Rater is who originally scored the image in the N10K image annotation subset - Pipeline: Raw images and disease annotations from the FHB image analysis pipeline - raterA: Raw images and manual images annotations of FHB disease by rater A - raterB: Raw images and manual images annotations of FHB disease by rater B - raterC: Raw images and manual images annotations of FHB disease by rater C - raterD: Raw images and manual images annotations of FHB disease by rater D - raterE: Raw images and manual images annotations of FHB disease by rater E 8_rover_20220713_StP_images_1-4: Raw rover images of Fusarium head blight on wheat from camera C and D for Saint Paul, MN on July 13th 2022 used for FHB image analysis pipeline evaluation 9_rover_20220718_StP_images_1-4: Raw rover images of Fusarium head blight on wheat from camera C and D for Saint Paul, MN on July 18th 2022 used for FHB image analysis pipeline evaluation 10_rover_20220720_StP_images_1-4: Raw rover images of Fusarium head blight on wheat from camera C and D for Saint Paul, MN on July 20th 2022 used for FHB image analysis pipeline evaluation 11_rover_20220728_Crk_images_1-4: Raw rover images of Fusarium head blight on wheat from camera C and D for Crookston, MN on July 28th 2022 used for FHB image analysis pipeline evaluation