[Tool] AI-Powered Phenotyping for Cassava Post-Harvest Quality

Post-harvest physiological deterioration (PPD) severely limits cassava shelf life and market value. We developed a deep-learning pipeline using YOLO foundation models and image segmentation tools to automatically detect and quantify PPD symptoms from simple RGB images. The system achieved over 80% detection accuracy, enabling rapid and scalable phenotyping compared with traditional manual scoring. By integrating artificial intelligence with plant breeding pipelines, this collaborative effort demonstrates how AI-driven phenotyping can accelerate the development of cassava varieties with improved storage stability and market value.