Detekcija lica dubokim neuronskim mrežama

undergraduate thesis

• Title (english): Face detection using deep neural networks
• Author: Vedran Ivanušić
• Mentor: Igor Sunday Pandžić (mentor)
• Committee member: Igor Sunday Pandžić (predsjednik povjerenstva)
• Committee member: Nenad Markuš (član povjerenstva)
• Committee member: Martina Manhart (član povjerenstva)
• Granter: University of Zagreb; Faculty of Electrical Engineering and Computing; Zagreb
• Defense date and country: 2017-09-12, Croatia
• Scientific / art field, discipline and subdiscipline: TECHNICAL SCIENCES; Computing

Abstract

Ovaj rad bavi se detekcijom lica uporabom dubokih neuronskih mreža. Rad opisuje način pripreme ulaznih podataka koja je od izrazite važnosti za ispravnost i kvalitetu detekcije. Navedeno uključuje normalizaciju ulaznih podataka, kao i predmeta koji je meta detekcije. Opisan je način rada konvolucijskih neuronskih mreža. Dan je pregled često korištenih aktivacijskih funkcija i kriterija pogreške. Kratko je opisan postupak učenja i optimizacije nad neuronskom mrežom. U završnom poglavlju, opisano je praktično ostvarenje detektora lica koristeći kaskadu konvolucijskih slojeva i koristeći detekciju pomoću posmačnog prozora.

Abstract (english)

This paper deals with face detection using deep neural networks. Paper describes the method
for preparation of the input data which is of significant importance for correct and quality detection. Aformentioned includes normalization of input data, as well as the object which is the target of detection. We describe the operation principles of convolutional neural networks. We make an overview of commonly used activation functions and error criteria. Shortly describing the learning method and optimization over neural network. In final chapter, we describe a practical implementation of a face detector using a cascade of neural network layers and a sliding window detection method.

• Keywords: Detekcija lica; neuronske mreže; konvolucijska mreža; Torch okvir
• Keywords (english): Face detection; neural networks; convolutional networks; Torch framework
• Language: croatian
• URN:NBN: urn:nbn:hr:168:392856
• 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: 2019-03-13 21:01:38