Machine Learning-based Object Recognition Technology for Bird Identification System

摘要

The object recognition system for machine learning has become widely known recently, and each algorithm of the object recognition system has its own advantages and disadvantages. The disadvantages are that the speed of identification is slow and the accuracy of identification is poor. The purpose of this paper is to identify birds through the YOLO and Custom Vision algorithms. When the user uploads bird photos to the two systems through a smart mobile device, the system will identify the objects according to the algorithm. It judges which bird it may be and, after a final comparison, the result with a higher percentage of accuracy is transmitted back to the user, thereby improving the accuracy of the identification system.