Medieval chant manuscripts digitized with new music recognition data
From Volpiano to Pixels: 30,000 Lines of OMR Ground Truth for Medieval Chant
Computer Vision and Pattern RecognitionInformation Retrieval
Summary
Medieval chant music is written in old, complex notation that is hard for computers to read. The authors created a large dataset linking images of medieval music manuscripts with their symbolic notation to help computers learn this. They trained a recognition model that can read musical symbols from these manuscripts much better than earlier efforts. This work helps convert old handwritten music into digital form for study and preservation.
What this means in practice
- •For digital music archivists: Improve the digital transcription of medieval chant manuscripts using a large labeled dataset and models trained on diverse notations.
- •For music software developers: Build better optical music recognition tools for historical notations using the supplied dataset and trained models to recognize different medieval music symbols.
Authors
Robert Klugseder
Abstract
Optical music recognition (OMR) for medieval chant lacks training data linking manuscript images to musical symbols across notational families. Using existing Volpiano transcriptions from the Cantus Database, we created 30,567 image-target pairs from ten sources in eleven exports, including 339 verse fragments. Their target sequences contain 1,952,143 tokens. The pipeline locates staff systems, infers clefs and maps known pitches to staff positions. Six detector families handle the different layouts and notations. Where heuristics failed, two small U-Nets learned staff lines and clef letters from manual annotations. Reviews of sampled export lines found zero to four pitch or clef errors, but used different error definitions and do not provide a uniform audit of the corpus. A CRNN trained from scratch achieved a symbol error rate of 9.6 % on 3,086 test lines. The earlier model, trained mainly on Austrian sources of the 15th and 16th centuries, scored 45.3 %. In three separate training runs, each excluding one manuscript, error rates on the excluded manuscript were 21.5-31.0 %. Model-assisted alignment recovered a further 2,686 lines.