Optical Music Recognition

Optical Music Recognition (OMR) aims to automatically transcribe musical notation from images into digital formats, significantly reducing the time and cost of manual transcription. Current research focuses on improving the accuracy and efficiency of OMR systems, particularly through the use of deep learning models like transformers and convolutional neural networks, often employing end-to-end architectures that process entire pages of music at once and addressing challenges posed by polyphonic music and complex layouts. These advancements are impacting music archiving, digital music libraries, and music education by enabling efficient digitization of historical scores and facilitating broader access to musical works.