Computer Aided Creative Design of Paper Packaging Based on Image Recognition in Graphic Design Teaching

Abstract

In the teaching of graphic design, making full use of computer aided design (CAD) can effectively improve the teaching level and enhance the teaching effect. In order to promote the innovation of graphic design teaching method and strengthen the application of CAD tools in paper packaging design, this article puts forward a packaging image recognition and integrity monitoring method based on improved Convolutional Neural Network (CNN), and realizes packaging creative design based on CAD modeling optimization. In order to train the sparse coding model end-to-end, the network regards the optimization process of solving the sparse minimization problem as a recursive network layer, and carries out end-to-end training together with the spatial pyramid pool layer and a deep CNN. The test results show that the paper packaging image recognition model in this article obviously improves the recognition rate. After the processing of this algorithm, the packaging image is effectively controlled by noise, the basic outline of the image is clearly visible, and the contrast of the image is improved to a certain extent, and the image is effectively enhanced, which can improve the creative design of paper packaging.