What is data annotation?

The most basic data label is the frame. For example, if the detection target is a car, the annotator needs to mark all the cars on a picture. The frame should completely block the circumscribed rectangle of the car. If the frame is not accurate, the machine may "learn badly". Another example is human gesture recognition, which contains 18 key points. Only trained annotators can master the labeling of these key points, and the labeled data can reach the standard of machine learning.

Automatic labeling technology is a technology formed on the basis of the development of computer graphics technology. The information stored in the database attribute table is mainly used to automatically label the theme features. When tagging, you can easily change the location, font, style, size and color of the tagging attribute by using the positive film of any field in the theme attribute table.

The main content of automatic annotation is map annotation. Map annotation is one of the basic contents of maps. Like other symbols on the map, annotation is also a kind of symbol, which plays a positioning role in many cases. It is an important way to transmit map information between cartographers and users. For example, points can be indicated according to the position and structure of notes, and the range of objects can be indicated according to the interval and arrangement trend of notes.