Primjena računalnog vida za raspoznavanje glazbenih nota

master's thesis

• Title (english): Application of computer vision for the recognition of musical notes
• Author: Dora Tomić
• Mentor: Mislav Grgić (mentor)
• Committee member: Mislav Grgić (predsjednik povjerenstva)
• Committee member: Sonja Grgić (član povjerenstva)
• Committee member: Jelena Božek (član povjerenstva)
• Granter: University of Zagreb; Faculty of Electrical Engineering and Computing; Zagreb
• Defense date and country: 2024-02-21, Croatia

Abstract

Optičko raspoznavanje glazbe je izazovno područje koje je na mnogo načina slično optičkom raspoznavanju teksta. Ipak, ono sa sobom nosi mnoge izazove s kojima se tradicionalni sustavi raspoznavanja bore. U novijim istraživanjima pokazalo se kako je pristup s kraja na kraj superioran kod raspoznavanja teksta pa je u ovom radu primijenjen za OMR. Implementirana su tri različita modela od kojih su posljednja dva osnovana na pristupu s kraja na kraj, dok je prvi prijelazna verzija između ovog i tradicionalnog pristupa. Također prvi model, budući da je jednostavniji, izgrađen je isključivo za monofone glazbene zapise, dok su druga dva proširena za rad i na polifonim zapisima. Modeli su na kraju međusobno uspoređeni.

Abstract (english)

Optical music recognition is a challenging field similar in many ways to optical text recognition. It brings, however, many challenges that traditional recognition systems struggle with. In recent research, it has been shown that the end-to-end approach is superior in text recognition, so in this paper, it was applied to OMR. Three different models have been implemented, the last two of which are based on the end-to-end approach, while the first is transitional version between this and traditional approach. Also, the first model, being simpler, was built exclusively for monophonic music scores, while the other two were expanded to work on polyphonic scores as well. In the end, the models were compared with each other.

• Keywords: optičko raspoznavanje glazbe; duboke neuronske mreže; monofoni notni zapisi; polifoni notni zapisi
• Keywords (english): optical music recognition; deep neural network; monophonic music scores; polyphonic music scores
• Language: croatian
• URN:NBN: urn:nbn:hr:168:040448
• Study programme: Title: Computing; Study programme type: university; Study level: graduate; Academic / professional title: magistar/magistra inženjer/inženjerka računarstva (magistar/magistra inženjer/inženjerka računarstva)
• Type of resource: Text
• File origin: Born digital
• Access conditions: Closed access
• Created on: 2024-03-28 07:55:48