The MONK system is a state-of-the-art handwriting recognition system, which is able to find handwritten words in images of historical documents. To do so, it needs examples. An example is a typed piece of text, a label, associated with an image. With each added label, MONK improves its pattern-recognition abilities. On the right, you find a quickstart manual.
Together with volunteers, we are currently training MONK in three different ways:
1. WORDS (R-TOOL)
With the R-Tool volunteers assign labels to individual words. The more often a word is labelled, the better MONK is able to find the word in our corpus. Labelling with the R-Tool leads to the creation of an index.
2. LINES (E-TOOL)
With the E-Tool volunteers transcribe lines. Transcribing lines establishes a basis for full text transcriptions.
MONK presents volunteers with hitlists of suggestions. If words are already known to MONK, then hitlists are a good way to increase the amount of labels.