One-Step Detection of Background, Staff Lines, and Symbols in Medieval Music Manuscripts with Convolutional Neural Networks.
Jorge Calvo-ZaragozaGabriel VigliensoniIchiro FujinagaPublished in: ISMIR (2017)
Keyphrases
- convolutional neural networks
- cultural heritage
- line detection
- automatic detection
- object detection
- convolutional network
- detection algorithm
- bounding box
- background segmentation
- computer vision
- complex background
- detection accuracy
- false positives
- straight line
- detection method
- foreground objects
- post processing
- detection rate
- false alarms
- foreground detection
- line segments
- hough transform