MAP Adaptation to Improve Optical Music Recognition of Early Music Documents Using Hidden Markov Models.
Laurent Pugin, John Burgoyne, Ichiro Fujinaga · 2007
Despite steady improvement in optical music recognition (OMR), early documents remain challenging because of the high variability in their contents. In this paper, we present an original approach using maximum a posteriori (MAP) adaptation to improve an OMR tool for early typographic prints dynamically based on hidden Markov models. Taking advantage of the fact that during the normal usage of any OMR tool, errors will be corrected, and thus ground-truth produced, the system can be adapted in real-time. We experimented with five 16th-century music prints using 250 pages of music and two procedures in applying MAP adaptation. With only a handful of pages, both recall and precision rates improved even when the baseline was above 95 percent. 1