A Lightweight and Effective Music Score Recognition on Mobile Phone

Recognition systems of scanned or printed music scores which have been implemented on personal computer received consideration of many scientists and achieved significant results in many years. A modern trend with music scores captured and played directly on mobile devices become more interesting to researchers. The limit of resource and the effects of illumination, distortion, inclination on input images are still challenges to these recognition systems. In this paper, we introduce a novel approach to recognize music scores captured by mobile camera. To reduce the complexity as well as computational time of the system, we groups all symbols extracted from music score into ten main classes. Each major class is applied to SVM to classify music symbols separately. The experimental results showed that our proposed method could be applied to real time applications and the performance is competitive to others.