Metoda za procjenu osvjetljenja scene

master's thesis

• Title (english): Method for Scene Illumination Estimation
• Author: Ilija Domislović
• Mentor: Sven Lončarić (mentor)
• Committee member: Sven Lončarić (predsjednik povjerenstva)
• Committee member: Marko Subašić (član povjerenstva)
• Committee member: Tomislav Petković (član povjerenstva)
• Granter: University of Zagreb; Faculty of Electrical Engineering and Computing; Zagreb
• Defense date and country: 2020-07-14, Croatia
• Scientific / art field, discipline and subdiscipline: TECHNICAL SCIENCES; Computing

Abstract

Ovaj diplomski rad bavi se problemom procjenu osvjetljenja u slici. Estimacije osvjetljenja samo je mali, ali važan dio računalnog vida. Eksplozija mobitela na tržištu stvorila je nastanak velikog broja slika koje je potrebno obraditi. U ovom diplomskom radu se rješava specifični problem procjene jednog iluminata i procjene dvoje iluminata. Za rješenje problema u diplomskom radu predložen je model koji koristi konvolucijsku neuronsku mrežu. Model se odvaja od drugih sličnih modela tako da ne stvara duboku neuronsku mrežu. Umjesto duboke, diplomski rad predstavlja široku neurološku mrežu koja koristi jednostavne informacije sa slike za procjenu iluminata.

Abstract (english)

This thesis tackles the problem of color constancy. Color constancy is only a small part of the ever-evolving field of computer vision. The explosion of cell phone on the market caused an influx of images that need to be processed. In this thesis the specific problem of single and dual illuminant estimation is tackled. A model using a novel convolutional neural network is proposed. The model separates itself from other similar models by not creating deep neural network. Instead, the thesis introduces a wide neural network that uses low level information from the image to do illuminant estimation.

• Keywords: color constancy; estimacije illuminata; estimacija više illuminata; konvolucijska neuronska mreža; k-means; tensorflow
• Keywords (english): color constancy; illuminant estimation; multi-illuminant estimation; convolutional neural networks; k-means; tensorflow
• Language: english
• URN:NBN: urn:nbn:hr:168:545592
• 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 12:07:26