Semantic Value Based Web Images Adaptation for Heterogeneous Client Networks*

Abstract

With the rapid development of computer and network, the Web sites with uniform contents cannot suit different networks and devices. Web content transcoding on proxy is a promising way to make Web documents adaptive to heterogeneous network conditions and browsing terminals, as it not only makes the servers maintain just one or a few versions of the content, but also reduces the data size delivered to clients. In this paper, we mainly focus on the adaptive transcoding of web images under bandwidth constraint, since compared with texts, images have much larger size, whose transcode can significantly reduce the totally delivered bytes. Firstly, the concept "semantic value" is proposed to represent the value of information provided by Web images. Then, image classification method is taken to assign the semantic value to images. Based on the semantic value assigning strategy, to optimally deliver the web contents over heterogeneous networks, an adaptive model is designed to decide proper transcoding way for each image so as to maintain as mush semantic value of transcoded Web Document as possible. Extensive experiments on 100 web pages containing 4201 images have shown the effectiveness and efficiency of our work.