Vrednovanje metoda za detekciju osoba u slikama

master's thesis

• Title (english): Evaluation of methods for detection of persons in images
• Author: Klara Marijan
• Mentor: Siniša Šegvić (mentor)
• Committee member: Siniša Šegvić (predsjednik povjerenstva)
• Committee member: Marko Čupić (član povjerenstva)
• Committee member: Jan Šnajder (član povjerenstva)
• Granter: University of Zagreb; Faculty of Electrical Engineering and Computing; Zagreb
• Defense date and country: 2022-07-14, Croatia
• Scientific / art field, discipline and subdiscipline: TECHNICAL SCIENCES; Computing

Abstract

U ovom radu obrazložen je problem detekcije objekata te princip funkcioniranja jednoprolaznih i dvoprolaznih detektora. Objašnjene su razlike među njima te su detaljno opisane arhitekture modela YOLOv5 i Faster R-CNN kao predstavnika obje vrste detektora. Uhodano je učenje navedenih mreža na vlastitom skupu podataka s ciljem razvoja modela koji detektira ljude te razlikuje djecu od odraslih osoba. Napravljena je usporedba rezultata koje daju oba modela te su oni navedeni na kraju rada.

Abstract (english)

This paper explains the problem of object detection and the principle of functioning of single-shot and two-shot detectors. The differences between them are explained in the paper and the architectures of YOLOv5 and Faster R-CNN models are described in detail, as they are representatives of both types of detectors. These models are trained on a custom dataset with the aim of developing a model that detects people and distinguishes children from adults. A comparison of the results given by both models is made and they are listed at the end of the paper.

• Keywords: detekcija objekata; jednoprolazni detektori; dvoprolazni detektori; YOLOv5; Faster R-CNN
• Keywords (english): object detection; single-shot detectors; two-shot detectors; YOLOv5; Faster R-CNN
• Language: croatian
• URN:NBN: urn:nbn:hr:168:182377
• 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: 2023-01-14 09:24:39