Application of image retrieval based on convolutional neural networks and Hu invariant moment algorithm in computer telecommunications

Abstract

Image recognition and retrieval is an important application field of digital image and it has a wide range of application scenarios in the computer telecommunications. Convolutional neural networks (CNN) is a kind of feed-forward neural networks which includes convolution calculation and which has a deep structure and it is one of the typical algorithms of deep learning. In recent years, it has become a highly efficient recognition algorithm which has been widely applied in such fields as pattern recognition and image processing. Its characteristics include few training parameters and strong adaptability. On the other hand, Hu invariant moment algorithm-a conventional algorithm, is also extensively used in various image processing fields due to its simple calculation and high efficiency. This paper analyzes the research background and significance of both CNN and Hu invariant moment algorithm, and introduces their research status. Besides, it also analyzes the results of color-, distance- and weight-based Hu invariant moment algorithm, and compares it with CNN to serve as a theoretical support for better achieving image classification, recognition and retrieval technology.