Title :
Segmentation method of offline mathematical symbols
Author :
Jakjoud, Widad ; Lazrek, Azzeddine
Author_Institution :
Dept. of Comput. Sci., Cadi Ayyad Univ., Marrakech, Morocco
Abstract :
This paper comes within the framework of designing a recognition system of mathematical symbols. Through this, we first recall the general process of recognition of mathematical symbols while focusing on the description of our method for segmentation of offline mathematical symbols. Our goal is to detect, extract and segment the different symbols of mathematical expression. This expression will be recognized later. The segmentation method which we proposed is based on the verification and/or rejection of alternatives in the different phases of segmentation to minimize doubt and to ensure a reversal in case of failure. Indeed, if several alternatives appear possible (doubt), the method adopts different solutions with weights. Making the decision would be delegated to the next levels of the process. The verification and rejection process can reduce the number of solutions to reduce the weighted doubt.
Keywords :
feature extraction; image segmentation; mathematics computing; object recognition; mathematical expression; offline mathematical symbol recognition system; offline mathematical symbol segmentation method; rejection process; verification process; Algorithm design and analysis; Databases; Feature extraction; Image segmentation; Noise; Pixel; Proposals; Early processing; Mathematical symbol; Recognition; Rejection; Segmentation; Verification;
Conference_Titel :
Multimedia Computing and Systems (ICMCS), 2011 International Conference on
Conference_Location :
Ouarzazate
Print_ISBN :
978-1-61284-730-6
DOI :
10.1109/ICMCS.2011.5945634