Researches advanced in image recognition

Abstract

Image recognition has always been a current study issue in the field of computer vision, which aims at training models to automatically predict the categories of objects contained in a given image. Early image recognition mainly relied on manual features, whose recognition accuracy is far from meeting the actual application needs. Thanks to the rapid development of deep learning technology, the image recognition algorithms based on convolutional neural network have made breakthroughs in recognition accuracy and speed. Nowadays, Image recognition has been widely used in various fields, such as security, medicine, aerospace. Through detailed literature analysis and investigation, this paper first introduces the representative image recognition algorithms; Secondly, we introduce the common data sets and evaluation indicators in the field of image recognition, and quantitatively compare the accuracy differences of different recognition are summarized and the future development of this field is predicted.