Machine Learning and Image Annotation

SuperAnnotate is focused on image annotation. We provide the backbone of building those outputs from AI, basically the data that is being fed to this machine learning algorithms that is being trained. Sometimes your model accuracy gets to, let's say, 90% accuracy. And in order to get to 92%, you need to know what to label next to get that accuracy. What are the edge cases, for example, where your model doesn't perform well? For example, your camera on the car can detect anything around when it's daylight, but in a nightlight, it will not work well. Or if it's a cloudy or a rainy, it's not going to work well. How