Here you can find additional packages for document analysis problems that are too specific to be included in the Gamera core distribution. These packages are provided as Gamera toolkits, which require a working Gamera installation. They can be installed like Gamera itself with the command
python setup.py build && sudo python setup.py install
Note that currently none of these toolkits has been ported to Python 3 and Gamera 4. These toolkits thus require Gamera 3 and Python 2.7.
Text document recognition
The OCR Toolkit is meant to help building optical character recognition (OCR) systems for standard text documents. It provides:
- a flexible mechanism for plugging in custom page segmentation algorithms
- heuristic rules for dealing with diacritics, and for disambiguation of commonly confused roman characters
- a ready-to-run python script ocr4gamera.py which acts as a very basic OCR-system
The GreekOCR Toolkit is an addon for the OCR Toolkit for polytonal (classical) Greek.
Staff line removal
The MusicStaves Toolkit provides algorithms and evaluation methods for staff line detection and removal, an important preprocessing step in Optical Music Recognition (OMR). The toolkit offers the following functionality:
- different algorithms for staff detection and removal
- an abstract interface for adding custom staff detection and removal algorithms
- evaluation algorithms for measuring staff removal quality
- deformation algorithms for creating test images for evaluation
Lute tablature recognition
The OTR Toolkit is a complete application for the recognition of historic lute tablature prints. It offers the following functionality:
- recognition of French, Italian, German lute tablature, and of modern guitar tablature (aka "spanish lute tablature")
- conversion of the tablature to music
Psaltic neume notation recognition
The Psaltiki Toolkit is a complete application for the recognition of the post 1800 neume based chant notation of the eastern church. It offers the following functionality:
- page layout analysis and baseline detection
- neume recognition and neume layout analysis
- generation of a machine readable output code
The Webcam Toolkit allows for automatically taking a snapshot photo from a webcam or document camera and converting it to the Gamera image data type. It provides
- a function for querying the attached video devices, and for selecting one of these
- a function returning a snapshot as a Gamera image
- wrappers around video drivers for MacOS X, Linux, and Windows
The FD Toolkit implements a wide variety of Fourier descriptors. These are useful features for shape recognition. In contrast to ordinary Fourier descriptors, the FD toolkit also provides desciptors that work on broken shapes.
MIS image file support
The MIS Support Toolkit adds reading support for the "Multiple Image Set" (MIS) image format that is used by the NIST Special Database 19 of the US National Institute of Standards and Technology, a widley used reference data set for OCR evaluation.