Your question is Improving Object Detection Accuracy. Take a moment with it on the right.
Talk me through your thinking if you like. When you're confident, submit your answer and I'll grade it like a real screen (7/10 or better passes).
What methods would you use to improve the accuracy of an object detection model?