|
Abstract:
|
In the last years, the recognition of handwritten mathematical formulae has recieved an increasing amount of attention in pattern recognition research. However,the diversity of approaches to the problem and the lack of a commerciallyviable system indicate that there is still much research to be done in this area.In this thesis, I will describe the previous work on a system for on-line handwrittenmathematical formulae recognition based on the structural constructionparadigm and two-dimensional grammars. In general, this approach can be successfullyused in the anaylysis of inputs composed of objects that exhibit rich structural relations. An important benefit of the structural construction is in nottreating symbols segmentation and structural anaylsis as two separate processeswhich allows the system to perform segmentation in the context of the whole formula structure, helping to solve arising ambiguities more reliably. We explore theopening provided by the polynomial complexity parsing algorithm and extend thegrammar by many new grammar production rules which made the system usefulfor formulae met in the real world. We propose several grammar extensionsto support a wide range of real mathematical formulae, as well as new featuresimplemented in the application. Our current approach can recognize functions,limits, derivatives, binomial coefficients, complex numbers and more. |