Image processing method

Segmentation: The image can be divided into meaningful blocks or areas of interest can be extracted by many methods, and detection methods can also be used, such as face detection and pedestrian detection. Segmentation generally requires high accuracy.

Recognition: Get the pictures or areas of interest in advance and classify them through machine learning, such as judging whether the object is an apple or a book. There are many applications.

Detection: detection has a clear purpose. In fact, it is a process of identifying samples, then training to get models, and finally directly matching the images.

Tracking: The simplest time-space matching of moving targets can be realized without using pattern recognition method. Of course, you can also use the method of detection and identification, which is slower.