A Real-Time Object Tracking Method

摘要

Artificial Intelligence is one of the most vital prospects in our current life; its applications vary from robotics, transport, houses, to industrial facilities and planes. While neural networks offer the ability to recognize and track new objects, it takes time during learning process especially deep neural networks. There are specific computer vision systems, which require high accuracy and high speed conditions, regardless the knowledge of tracked object. Such as a tracking system for a measurement instrument which scans a certain area and follows the location of object(s). In this paper, we introduce a new tracking method for one object in real time system, based on a theory, building a new powerful algorithm to track the center of object in sub-pixel accuracy.