Title: Estimating similarity of rich internet pages using visual information

Authors: Zhen Xu; James Miller

Addresses: Department of Electrical and Computer Engineering, University of Alberta, 11-203 Donadeo Innovation Centre for Engineering, 9211-116 Street NW, Edmonton, Alberta, T6G 1H9, Canada ' Department of Electrical and Computer Engineering, University of Alberta, 11-203 Donadeo Innovation Centre for Engineering, 9211-116 Street NW, Edmonton, Alberta, T6G 1H9, Canada

Abstract: Traditional text-based web page similarity measures fail to handle rich-information-embedded modern web pages. Current approaches regard web pages as either DOM trees or images. However, the former only focuses on the web page structure, while the latter ignores the inner connections among different web page features. Therefore, they are not suitable for modern web pages. Hence, the idea of a block tree is introduced, which contains both structural and visual information of web pages. A visual similarity metric is proposed as the edit distance between two block trees. Finally, an experiment is undertaken, by cross-comparing 500 web pages, illustrating that the model appears to be highly accurate, empirically demonstrating that the metric is highly promising.

Keywords: block tree; gestalt laws of grouping; normalised compression distance; NCD; tree edit distance; TED; web page classification.

DOI: 10.1504/IJWET.2017.086415

International Journal of Web Engineering and Technology, 2017 Vol.12 No.2, pp.97 - 119

Published online: 10 Sep 2017 *

Full-text access for editors Full-text access for subscribers Purchase this article Comment on this article