Algorithm Analysis of Clothing Classification Based on Neural Network

Abstract

With the rapid development of Internet e-commerce, the online transaction volume of clothing has increased day by day, and the importance of clothing images in transactions has also increased. However, there are many clothing categories and different classification standards. It is difficult for consumers and e-commerce merchants to unify the description of clothing categories, which can easily lead to a poor clothing shopping experience. Neural network has an excellent list in the field of computer vision, which can effectively classify clothing. The purpose of this article is to study the algorithm analysis of clothing classification based on neural network. Starting from the neural network, this paper proposes a clothing image classification algorithm based on a multi-task convolutional neural network (Convolutional Neural Network, CNN). Through hierarchical classification data combined with multi-task technology, the basic structure of the network model is not changed. The accuracy of clothing image classification improves the network’s ability to express refined clothing categories. This paper proposes a clothing classification algorithm based on the feature fusion of Hu invariant matrix and CNN network. The feature fusion of the features extracted by the convolutional neural network is initially explored, the information gain of the feature is calculated, and the shape feature is used to eliminate the feature with less information gain. This paper also designs a clothing classification system based on neural network to realize the recognition, detection and classification of clothing images. The experimental results show that the clothing classification accuracy rates under the four combined tasks are 93.54%, 89.26%, 92.14%, 95.66%, and 93.54%, respectively. It can be seen that the model based on convolutional neural network can further improve the accuracy of clothing classification.