A rotation-robust script identification based on BEMD and LBP

Script identification is very important part to develop the scripts OCR systems. In this paper, we proposed a new algorithm for script identification based on the global and local texture of document images. The bidimensional empirical mode decomposition (BEMD) method is used to decompose the image to some components (IMFs) and then the Local Binary Patterns (LBP) method is used to detect the features. Experiments shown the recognition rate based on BEMD-LBP is as well as the LBPV and wavelet based energy feature in 0 angles. At the same time, for the different angles rotation script, the BEMD-LBP feature present some robust adaptive to the rotation script.