An interference recognition algorithm for auditing engineering drawings based on optimised CNN
by Jing Zhang; Jie Li; Ailv Zhu; Zining Li
International Journal of Information and Communication Technology (IJICT), Vol. 22, No. 3, 2023

Abstract: Engineering drawing is one of the main objects of audit work. At present, the use of computer automatic identification technology to digitise engineering drawings has become the main method to solve the low efficiency of traditional audit work. However, few reports on solutions for sample attacks based on engineering drawings. An interference recognition algorithm for auditing engineering drawings based on optimised convolutional neural networks (CNN) is proposed in this paper. The purpose is to be able to proactively identify whether the engineering drawings contain adversarial samples before carrying out the audit of engineering drawings. The specific methods are as follows: ten types of images that may be embedded in engineering drawings are considered firstly; then, the LeNet and VGG networks were improved for these ten types of images, respectively. And the network model with high recognition accuracy against sample images is trained and generated. Finally, compared with the other three networks, the experimental results show that the accuracy of IVGG is improved on average 4.44% than the other three methods.

Online publication date: Mon, 03-Apr-2023

The full text of this article is only available to individual subscribers or to users at subscribing institutions.

 
Existing subscribers:
Go to Inderscience Online Journals to access the Full Text of this article.

Pay per view:
If you are not a subscriber and you just want to read the full contents of this article, buy online access here.

Complimentary Subscribers, Editors or Members of the Editorial Board of the International Journal of Information and Communication Technology (IJICT):
Login with your Inderscience username and password:

Username:        Password:          Forgotten your password?

Want to subscribe?
A subscription gives you complete access to all articles in the current issue, as well as to all articles in the previous three years (where applicable). See our Orders page to subscribe.

If you still need assistance, please email subs@inderscience.com