82% accuracy marks early success for AI leaf-damage tool

Stony Brook University student Amy Chen is building an artificial intelligence system to detect environmental stress in vegetation. Her initial model achieved 82% test accuracy in distinguishing between damaged and healthy leaves, a benchmark that recently earned her the title of Researcher of the Month for October 2026 from the Undergraduate Research, Creativity and Artistic Activity (URECA) program.

Chen, an Information Systems major graduating in 2027, began this work in March 2025 under the guidance of Guanyu Huang. Supported by the SUNY SOAR undergraduate research program, she helped establish an ozone bioindicator garden. Her primary tasks involved photographing plants and systematically tracking the progression of leaf damage caused by ozone exposure.

These field observations served as the training data for the computer vision project. While the first iteration successfully classified leaves as either damaged or healthy, it lacked the precision to quantify the extent of that damage. To address this gap, the research team is currently expanding the dataset by collecting additional photographs from other ozone gardens to refine the model's measurement capabilities.

Ahmad Esmaili, senior lecturer and Information Systems Undergraduate Program Director, noted that the project demonstrates how an Information Systems student can apply data science to environmental questions. Chen presented her findings at the American Meteorological Society’s annual meeting in January 2026, where she received a best student poster award. She also presented her work at the National Ozone Garden Network Conference in August 2026.

Chen is now preparing a manuscript with her research group. She will incorporate this research into her WISE senior thesis. As a WISE honors student with a minor in Environmental Studies, she is integrating her technical skills in information systems with her focus on environmental science.