Towards a foundational model for recognising diastematic Gregorian notation

2026-06-30

AI Summary

This work addresses the limited generalizability of existing optical music recognition methods for Gregorian chant notation, which stems from inconsistent dataset encodings. To overcome this challenge, the authors propose S-GABC, a unified encoding framework that integrates four heterogeneous datasets for the first time and enables the development of the first end-to-end foundational model specifically designed for diastematic Gregorian chant notation. By jointly training on all four datasets under this standardized representation, the proposed method achieves state-of-the-art performance across every benchmark dataset, substantially improving model generalization. This approach offers a universal and scalable solution for Gregorian chant recognition, paving the way for more robust and interoperable systems in historical music notation analysis.

Abstract

Optical recognition of Gregorian notation has recently been attempted with end-to-end methods, with four datasets introduced. However, each of these datasets is in a different encoding. We design a common encoding based on the S-GABC proposal, convert all four datasets to this common encoding, and train a shared end-to-end foundational model for diastematic Gregorian notation that establishes a new state of the art across all four datasets.

Problem

• Gregorian notation
• optical recognition
• dataset encoding
• foundational model
• diastematic notation

Innovation

• foundational model
• diastematic Gregorian notation
• optical music recognition
• unified encoding
• S-GABC