Cassava breeders collect several key measurements during the breeding cycle to assess important traits such as Cassava Brown Streak Disease (CBSD) spread on cassava roots (necrosis), whitefly count, and Postharvest Physiological Deterioration (PPD), which affects the shelf life and usability of cassava across different genotypes. These measurements are essential for breeders to make informed decisions on which genotypes to advance through the breeding pipeline.
Currently, these assessments rely on visual inspection and manual estimation, which are often subjective, prone to human error, and limited in throughput due to the complexity of the tasks. This project introduces Machine Learning tools designed to support cassava breeders with accurate, consistent, and high-throughput measurements, enabling data-driven decision-making and improving the efficiency and precision of the cassava breeding process.