Research And Implementation Of Web Page Adaptation On Small Screen Devices Based On Web Page Segmentation And Summarization

With the popularization of handheld devices, application mode to obtain web information using handheld devices has developed rapidly. Currently most web pages are designed for PC's screens commonly. Because of handheld devices' small-sized screens, they lead to awkward operation and inefficient web browsing. Therefore how to improve the efficiency of web viewing on small screen devices and reduce the burden of user operation has become a popular research topic nowadays.This paper researches on web page adaptation on small screen devices and proposes a new solution based on web page segmentation and summarization techniques. We improve the two techniques to make them solve the web viewing on small display more efficiently. In order to have better adaptation, our solution also analyzes the device profile to obtain different summary according to different screen size. Main works are as follows:In regard to page segmentation, an automatic page segmentation algorithm using multi-cues is proposed based on analyzing current techniques and their shortage. Different with existing algorithms using only one cue, both non-visual cues and visual cues are taken into account for following human perception. With these semantic cues, we analyze the structure of an existing web page and split it into small units, and then use pattern discovery to adjust segmentation results. The algorithm can be used in many applications, not just for the one in this paper. Experiments show better accuracy and more flexibility of our algorithm over existing ones.We also explore the techniques for web page summarization. The web page summarization algorithm based on display size is presented by combining the summarization techniques and application environment of small screen device. The key point of this algorithm is screen adaptation, that is to say, the summary is different according to the screen size. Except for traditional methods for summarization, the semantic information hidden in Semi-Structure information is also utilized to improve the quality of summary.What's more, the prototype system on proxy sever is designed and implemented. Main modules are explained in details and experiments are designed aiming at the two presented algorithms.