Optical recognition tool extracts notes from string quartet scores
An Optical Music Recognition Tool for String Quartets
Computer Vision and Pattern Recognition
Summary
Reading music from printed sheets is tricky, especially for groups like string quartets whose music can be complex. The paper presents a computer system that looks at scanned or digital music sheets and identifies notes, rhythms, and other musical marks accurately. It uses special artificial intelligence methods to find and classify these musical symbols, then checks that rhythms make sense according to the time signature rules. The system was tested and shown to work well, and its code and data are publicly shared.
What this means in practice
- •For music software developers: Integrate precise recognition of string quartet sheet music into digital music editing and analysis tools.
- •For digital archivists: Digitize and archive string quartet scores with accurate musical detail extraction from PDF sources.
Authors
Jin-Shiang Hu, Bo-Yuan Cheng, Li Su
Abstract
This paper presents an optical music recognition (OMR) system tailored for PDF scores of string quartet pieces. The approach begins with a SegNet-based segmentation model to localize musical regions, followed by convolutional neural networks and image processing techniques to classify and extract musical symbols. To improve rhythmic precision, dynamic programming methods enforce time signature constraints and explicitly handle tuplets. Evaluation using a longest common subsequence metric across multiple feature matrices demonstrates robust accuracy in recognizing pitch, onset, duration, and accidentals on a benchmark dataset. Code and data are available at https://github.com/lattellie/25-omr.