Detekcija i lokalizacija osoba u slikama korištenjem konvolucijskih neuronskih mreža

master's thesis

• Title (english): Detection and Localization of Persons in Images Using Convolutional Neural Networks
• Author: Josip Mandić
• Mentor: Tomislav Hrkać (mentor)
• Committee member: Tomislav Hrkać (predsjednik povjerenstva)
• Committee member: Zoran Kalafatić (član povjerenstva)
• Committee member: Alan Jović (član povjerenstva)
• Granter: University of Zagreb; Faculty of Electrical Engineering and Computing; Zagreb
• Defense date and country: 2017-07-06, Croatia
• Scientific / art field, discipline and subdiscipline: TECHNICAL SCIENCES; Computing

Abstract

Rad se bavi izradom rješenja za detekciju i lokalizaciju osoba na slikama uporabom konvolucijskih neuronskih mreža. Na početku se uvodi u područje računalnog vida i strojnog učenja. Opisuju se osnovni pojmovi neuronskih mreža. U okviru ovog rada, proučeni su postupci učenja modela konvolucijskih neuronskih mreža. Iskorišten je naučeni model za klasifikaciju VGG-16. Napravljene su promjene u strukturi VGG modela, te je prilagođen rješavanju problema ovog rada. Prikazano je impelementirano rješenje te su opisane korištene metode i tehnike. Zatim su prikazani razultati treniranja modela i validacije.

Abstract (english)

This paper describes solution for the person detection and localization in images using the convolutional neural networks. At the beginning, we are introduced into the field of computer vision and machine learning. The basic concepts of neural networks are described. In this study, CNN model learning methods have been studied. Pre-trained model VGG-16 has been used. Original VGG model has been changed and adapted to solve our problem. An implemented solution is shown and the used methods and techniques are described. At the end results of the model training and validation are shown.

• Keywords: Računalni vid; strojno učenje; neuronske mreže; konvolucijske neuronske mreže; tensorflow; keras; python
• Keywords (english): Computer vision; machine learning; neural networks; convolutional neural networks; tensorflow; keras; python
• Language: croatian
• URN:NBN: urn:nbn:hr:168:292846
• 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: 2019-03-13 22:36:02