Sustav za raspoznavanje nota u slikama glazbenih notnih zapisa

master's thesis

• Title (english): System for Musical Notes Recognition in Sheet Music Images
• Author: Ivan Cindrić
• Mentor: Tomislav Hrkać (mentor)
• Committee member: Tomislav Hrkać (predsjednik povjerenstva)
• Committee member: Zoran Kalafatić (član povjerenstva)
• Committee member: Jan Šnajder (član povjerenstva)
• Granter: University of Zagreb; Faculty of Electrical Engineering and Computing; Zagreb
• Defense date and country: 2021-07-02, Croatia
• Scientific / art field, discipline and subdiscipline: TECHNICAL SCIENCES; Computing

Abstract

Tema je program koji prepoznaje i čita notni zapis.
Ulaz u program je slika notnog zapisa (.jpg ili .png format).
Korištenjem računalnog vida (OpenCV biblioteka) i neuronskih mreža, slika je prvo na redove, a zatim se iz pojedinog reda izvlače zasebni elementi(note, ključ, modulacija, itd...). Ulaz u mrežu je jedan element notnog zapisa(npr. jedna nota). Trajanje i vrijednosti note se određuju pomoću dvije neuronske mreže. Jedna mreža služi za prepoznavanje visine tona. Druga mreža notama pridružuje trajanje. Kombiniranjem rezultata tih dviju mreža stvara se novi notni zapis. Izlaz programa je rekonstruirana pjesma spremljena u .midi formatu.

Abstract (english)

The topic is a program that recognizes and reads sheet music.
The input into a program is a sheet music image (in .jpg or .png format).
By using computer vision (OpenCV library) and neural networks, the image is first segmented into individual lines, and then single elements are extracted from those lines (notes, clefs, key signatures, etc...). Input into the neural network is one element of the sheet music (for example, one note).
Program consists of two neural networks. One network is used to identify the pitch of a note (if the element is a note) and to identify the meaning of the rest of the elements (for example, treble clef or bass clef, time signature , key signature...). The other network associates notes with their duration. By combining those two networks, a new sheet music is created. The program output is a reconstructed song saved in .midi format.

• Keywords: prepoznavanje nota; računalni vid; neuronske mreže; yolo V3
• Keywords (english): note recognition; computer vision; neural networks; yolo V3
• Language: croatian
• URN:NBN: urn:nbn:hr:168:284655
• 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: 2023-01-13 09:51:28