Mole is intended to be an OCR system for recognition of printed or in best case handwritten text data. Currently some functions, in particular the segmentation algorithm, are very simple and may be insufficient at data with low image quality or complex structure. Mole is designed as a modular system under C++ and wxWidgets. All core functionalities are established as plugins which can be improved or substituted easily at compile time. Mole is more like an OCR framework with core components which can be substituted to fit individual needs. Supported modules include segmentation, feature extraction, data handling and classification. (Further steps such as pre- and post processing could be supported but should be done by more appropriate programs better suited for this purpose such as GIMP or a spell checker as Aspell). Mole comes with a prepared parameter set, so it should be capable to work from startup on printed data in an acceptable way (but do not expect to much, mole is experimental). Mole offers methods to increase recognition rates by learning and parameter modification, but I hope also without such a step it may work often.
