Enhancing Recognition of Historical Musical Pieces with Synthetic and Composed Images

Handwritten Music Recognition (HMR) poses the problem of transcribing historical musical pieces from digital image to text. The vast number of untranscribed pieces, together with the scarcity of manually annotated training data renders the manual transcription impractical. Historical musical pieces of particular interest are those dating back to the XVth century and earlier, available only in their original manuscripts. Current state-of-the-art approaches leverage Convolutional and Recurrent Neural Networks (CRNN) due to their effectiveness in processing information without relying on extensive datasets. This paper addresses the data scarcity challenge in HMR by proposing two approaches. Firstly, the utilization of synthetic images to augment the training data, leveraging its successful applications in Handwritten Text Recognition (HTR). Secondly, the paper advocates for image composition, combining the images from a manuscript page to mitigate the contextual limitations associated with single-line processing. Despite challenges observed in traditional HTR models, the regularity found in historical musical document layouts enhances the suitability of image composition for HMR. These approaches allow us to develop a system that can take advantage of additional samples and contextual information to improve the recognition capabilities of the HMR models. Results obtained show a relative improvement when working with synthetic images and a substantial improvement when image composition is considered.