I was working on a project to detect defects in LCD array using object detection model. Good accuracy models are of utmost importance as it can:
* Significantly reduce worker's labelling time
* Reduce number of arrays needed to be fixed
* Most importantly, reduce cost
Achievements:
* Using YOLOv5 model with above 90% accuracy
* Using state-of-the-art YOLOv7 to compare with the renowned YOLOv5
* Design and develop user interface for ease of use
* Implement auto label and model retrain feature onto UI