AI Weed Detection Pipeline — Data Annotation & Labeling (Final Year Project)
As part of my final year project, I audited and relabeled complex image datasets to ensure accuracy for AI weed detection. I optimized the dataset for high-fidelity AI training by reassessing prior labels and validating model performance. This process involved documenting labeling decisions and using confusion matrices to analyze labeling outcomes. • Conducted dataset relabeling for improved weed detection accuracy. • Utilized object detection methods to prepare images for YOLOv11 training. • Recorded labeling choices and evaluation steps in technical documentation. • Evaluated model performance to ensure annotation quality was maintained.