Research on Flower Image Classification Algorithm Based on Convolutional Neural Network

Abstract

The fast progress of deep learning makes Convolutional Neural Network (CNN) emerges at the historic moment, and as an important achievement, it has been extensively used in all sorts of fields. Compared with traditional machine learning, CNN has more advantages, on the one hand, it has more hidden layers and complex network structure, and on the other hand, it has a stronger ability of feature learning and feature expression. With the fast progress of computer technology, the application research of fast and accurate recognition and classification of flowers by obtaining flower images through mobile devices has received extensive attention. The flowers images collected under natural conditions have large background interference, and it is difficult to recognize flowers because of their inter-class similarity and intra-class diversity. Therefore, in view of the lack of flower image data and low classification accuracy, t he experi m ent sorted out the data sets of four kinds of flowers, and used the CNN to classify the images. Compared with the traditional approaches, the classification precision can be largely enhanced.