Abstract: A Web page having display elements such as a headline, a story body, subheads, and links to articles is obtained and rendered internally to obtain a position of each display element based on draw data. Each display element is classified into several clusters according to its position. Next, clusters having the same character attributes are classified as a group. A group having a high average of the number of characters within each of its clusters is determined as the story body and a group having a low average is determined as the headline. Then, individual pages including the story body and a top page including the headline, the subheads, and links to the story body pages are created. Therefore, the Web page is reconstructed as Web pages suitable for browsing in low-resolution display environments.

Abstract: A Web page including as its display elements a headline and a body of its story, subheads and links to articles belonging to the subheads is obtained and is rendered internally so as to obtain a position of each display element on the basis of draw data. Each display element is classified into several clusters in accordance with the obtained position of the display element and layout features of the individual clusters are detected so as to discriminate clusters of the headline and the body of story in accordance with a result of the feature detection. Next, clusters having same character attribute are collected to be a group, and a group having a high average of the number of characters within each cluster included in each group is determined as the body of story and a group having a low average as the headline. Then, individual pages of the body-of-story/article are created and a top page including the headline and subheads provided with links to these body-of-story/article pages is created.