Plant Pathology Challenge 2021

Overview

Foliar diseases are major problems for the fruit quality and overall productivity of apple orchards. The current disease diagnosis in apples is based on manual scouting by humans, which is time-consuming and expensive. Computer-vision based models are promising way to efficiently perform automatic disease identification. However, the major challenges for computer-vision based accurate disease identification are variation in visual symptoms of a single disease across different apple cultivars due to differences in leaf color, leaf morphology, age of infected tissues, non-uniform image background, and different light illumination during imaging. We have collected high-quality RGB images of foliar diseases of apples and created an expert-annotated disease dataset. This dataset reflects the real world scenario by representing non-homogeneous background of leaf images taken at different maturity stage and at different times of day under different camera focal settings. The challenge this year is to go above and beyond the significant progress of last year in which more than 1,300 teams participated with more than 22,000 ML models. The dataset for this competition has been significantly increased from 3,651 last year to more than 23,000 and we have added two additional disease categories this year. The main objectives of the competition is to develop machine learning-based models; 1) To accurately classify the given leaf image from the test dataset to particular disease category, 2) To develop an algorithm for accurate estimation of disease severity of each image/disease category; and 3) To identify individual disease from multiple disease symptoms on a single leaf image.

Competition

Start Date - March 15, 2021

End Date - May 26, 2021

Kaggle URL - https://www.kaggle.com/c/plant-pathology-2021-fgvc8

Organizers

Awais Khan and Ranjita Thapa (Plant Pathology and Plant-Microbe Biology Section, Cornell AgriTech, Cornell University) Serge Belongie, Noah Snavely, Qianqian Wang (Department of Computer Science, Cornell Tech, Cornell University)

Acknowledgements

We acknowledge financial support from Cornell Initiative for Digital Agriculture (CIDA) and special thanks to Elizabeth Marie Tee for providing some images of powdery mildew disease symptoms.