Book inner boundary extraction with modified active shape model

摘要

This paper proposes one modified active shape model (MASM) method to extract book inner boundaries in the scanned book images. It assumes that both pages are included in the scanned image and the book page corners are provided. The MASM method introduces the "book shape" idea and represents the book shape as one set of landmark points which are sampled from book boundary. It utilizes one iterative part-to-whole strategy to recognize the landmark points of inner boundary: at the part level, dynamic programming is used to incorporate into the specific properties of book boundary to locate the landmark points for each piece of boundary; at the whole level, one book shape model is used to constrain the shape composed by the landmarks. Finally, it extracts the inner boundary according to the located landmark points. In addition, one "expanding training shapes" strategy is proposed to learn one shape model with better generalization ability. The proposed method is evaluated on the collected datasets and shows the impressive result.