Understanding the page logical structure
Luigi Cinque, L. Forino, STEFANO LEVIALDI, Luca Lombardi, S. Tanimoto · 2003
We present an algorithm to segment and interpret technical journal pages. Our method has some advantages: it performs a fine segmentation and classification, after a preliminary step of page cleaning, and analyses the spatial disposition of the classified blocks in the page, which is generally not included in other methods. Moreover, there is a choice of working on the source page at two different resolution levels, allowing memory space to be saved and efficiency to be increased.