Image classification based on CNN with three different networks
Sun Wenyan · Applied and Computational Engineering · 2023
Image classification refers to classifying images based on the different features reflected by the information in each image. Image classification is a fundamental issue in computer vision and has important significance. It has a wide range of applications, such as autonomous driving, face recognition, image retrieval, and other fields. This article first gives a bird’s-eye view of the development of image classification and briefly introduces the factors that affect the accuracy of convolutional neural networks. Experiments and comparative analysis are conducted on the effects of convolutional layer numbers and optimizers on the accuracy of convolutional neural networks. Three convolutional neural networks with different convolutional layer numbers are built, their structural design is introduced in detail, and their structural diagrams were given. Different models are used to classify images from the CIFAR 10 dataset. The experimental results show that with the continuous increase of the convolutional layer, the accuracy rate is continually improving, but the program running time is continually increasing. In addition, using different optimizers can lead to changes in accuracy.