Usporedba klasičnih metoda za detekciju lica na slici i metoda utemeljenih na dubokom učenju

undergraduate thesis

• Title (english): Comparison of classical methods for face detection in image and methods based on deep learning
• Author: Luka Šantek
• Mentor: Sonja Grgić (mentor)
• Committee member: Sonja Grgić (predsjednik povjerenstva)
• Committee member: Mislav Grgić (član povjerenstva)
• Committee member: Tomislav Kos (član povjerenstva)
• Granter: University of Zagreb; Faculty of Electrical Engineering and Computing; Zagreb
• Defense date and country: 2023-06-14, Croatia
• Scientific / art field, discipline and subdiscipline: TECHNICAL SCIENCES; Computing

Abstract

Detekcija lica na slici važan je problem računalnog vida široke praktične primjene. Postoje dva osnovna pristupa problemu – prvi podrazumijeva klasične metode, čiji se rad temelji na analizi značajki, dok se drugi, noviji pristup temelji na umjetnim neuronskim mrežama i dubokom učenju. Objašnjavaju se najvažniji koncepti i pojmovi vezani uz općeniti problem detekcije lica na slici. Daje se pregled teorije ključne za razumijevanje principa rada svake od metoda, navode specifičnosti te objašnjava konkretna algoritamska implementacija. Provodi se testiranje na vlastitom skupu podataka te vrši međusobna usporedba rezultata, kao i usporedba s analizama iz postojeće stručne literature. Izvode se odgovarajući zaključci te navode glavne prednosti i nedostatci za oba pristupa.

Abstract (english)

Face detection is an important problem of computer vision with wide range of practical applications. There are two basic approaches to the problem - the first involves classical methods, whose work is based on feature analysis, while the second, newer approach is based on artificial neural networks and deep learning. The most important concepts and terms related to the general problem of face detection are explained. An overview of the theory crucial for understanding the working principle of each of the methods is given and a specific algorithmic implementation is explained. Testing is performed on own dataset and the results are compared with each other, as well as with analysis from the existing professional literature. Appropriate conclusions are drawn and the main advantages and disadvantages of both approaches are stated.

• Keywords: detekcija lica; računalni vid; strojno učenje; duboko učenje; neuronske mreže; konvolucijske neuronske mreže; umjetna inteligencija; kaskadni klasifikatori; klasične metode; analiza značajki
• Keywords (english): face detection; computer vision; machine learning; deep learning; neural networks; convolutional neural networks; artificial intelligence; cascade classifier; classical methods; feature analysis
• Language: croatian
• URN:NBN: urn:nbn:hr:168:492767
• Study programme: Title: Computing; Study programme type: university; Study level: undergraduate; Academic / professional title: sveučilišni/a prvostupnik/ prvostupnica (baccalaureus/baccalaurea) inženjer/inženjerka računarstva (sveučilišni/a prvostupnik/ prvostupnica (baccalaureus/baccalaurea) inženjer/inženjerka računarstva)
• Type of resource: Text
• File origin: Born digital
• Access conditions: Closed access
• Created on: 2023-07-13 10:34:48