Image Classification Research
I led a team that labeled over 2,000 images/documents to train an object detection model. We implemented a quality assurance workflow that improved data accuracy by 9%, which directly contributed to the final model achieving an F1-score of 0.98