Design and Implementation of Aerospace Arisa Fashion Design CAD System Based on Deep Convolution Neural Network

Abstract

The application of garment computer aided design (CAD) technology is not only a sign reflecting the scale and benefit of garment enterprises, but also a technical guarantee for garment enterprises to improve their competitiveness. Its basic principle is to construct a multi-level exploratory neural network model, use the image data represented by existing classification to supervise the network model in advance, and to achieve the function of accurately identifying unknown types of images. In this text, a multi-feature fusion image feature extraction method based on Deep learning (DL) is proposed. The DCNN model based on GA optimization is used to extract clothing style information from clothing images, and the multi-features obtained from DCNN model with transfer learning are combined with image texture features, and then input into XGBoost framework for learning, so as to improve the accuracy of clothing layout method. The experimental results show that the model accuracy of our method is more accurate than traditional neural networks, and the modeling accuracy is obviously improved. The algorithm uses artificial neural network technology to simulate the experience and technology of pattern designers, and provides support for the design and implementation of aerospace Arisa fashion design CAD system.