AI-ENGAGE: Image-based phenotyping and multi-omic prediction schemes for horticultural crop improvement

About This Grant

Fruit and vegetable crops are essential to a healthy diet, supplying nutrients often missing from staple grains like corn and wheat. Their high market value also offers important economic opportunities for farmers. Developing new varieties with improved disease resistance, yield, and quality benefits both producers and consumers. However, characteristics that set these crops apart, such as their perishability, also make them harder to breed. A major challenge is collecting reliable data on yield and quality, a process that is often slow, costly, and labor-intensive. This project addresses that challenge by creating faster, more precise ways to evaluate and select for these traits in three of the world’s most important horticultural crops: tomato, onion, and strawberry. It brings together researchers from the U.S., India, Japan, and Australia to apply advanced tools in imaging, machine learning, and genomics to support the development of productive, high-quality varieties that meet the needs of both growers and consumers. This project develops methods for the high-throughput, non-destructive evaluation of yield and quality using both RGB and spectral imaging. To accomplish this, data are collected from both handheld and autonomous devices and fed into deep learning-based image segmentation models to measure traits such as fruit count, size, and shape in tomato and strawberry as well as bulb shape in onion. In addition, the project investigates the ability of models incorporating high-dimensional biological data, including hyperspectral, genomic, and transcriptomic features, to predict complex traits in these crops. The research team also combines 3D modeling and gene expression data to understand and forecast growth in strawberry plants. In parallel, the project fosters international collaboration and capacity-building through research exchanges, workshops, and training opportunities focused on the use of modern phenotyping and predictive tools in plant breeding. This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.