Razvoj sustava za prepoznavanje geometrijskih tijela korištenjem YOLOv8 modela

master's thesis

• Title (english): Development of a system for recognizing geometric bodies using the YOLOv8 model
• Author: Filip Zadro
• Mentor: Tomislav Stipančić (mentor)
• Granter: University of Zagreb; Faculty of Mechanical Engineering and Naval Architecture; Zagreb
• Defense date and country: 2024-07-15, Croatia
• Scientific / art field, discipline and subdiscipline: TECHNICAL SCIENCES; Mechanical Engineering

Abstract

Računalni vid se koristi u različite svrhe, uključujući prepoznavanje, detekciju, klasifikaciju. Prilikom treniranja neuronske mreže koriste se različite metode umjetne inteligencije, koje koriste prikupljene podatke za učenje modela. Cilj ovog diplomskog rada je razviti sustav prepoznavanje geometrijskih tijela za društvenu igru Dungeons&Dragons (D&D) koristeći YOLOv8 paket. Potrebno je razviti neuronsku mrežu. prikupiti potrebne podatke te prikupljene podatke je potrebno obilježiti i klasificirati, nakon toga provesti treniranje te izvršiti analizu rezultata. Skup podataka se sastoji od „kockica“ D4, D6, D8, D10, D12, D20 i D100. Treniranje modela se provodi s varijabilnim skupom podataka, tako da će se skup podataka povećavati novim podacima nakon svakog treninga, s ciljem praćenja poboljšavanja neuronske mreže.

Abstract (english)

Computer vision is used for various purposes, including recognition, detection, and classification. During the training of a neural network, different artificial intelligence methods are utilized, using collected data to train the model. The aim of this thesis is to develop a geometric shape recognition system for the board game Dungeons & Dragons (D&D) using the YOLOv8 package. It is necessary to develop a neural network, collect the required data, label and classify the collected data, conduct the training, and perform an analysis of the results. The dataset consists of "dice" D4, D6, D8, D10, D12, D20, and D100. The model training is conducted with a variable dataset, in such a way that the dataset will be increased with new data after each training session, with the goal of monitoring the improvement of the neural network.

• Keywords: Računalni vid; Neuronska mreža; YOLOv8; Klasifikacija
• Keywords (english): Computer vision; Neural network; YOLOv8; Classification
• Language: croatian
• URN:NBN: urn:nbn:hr:235:397407
• Study programme: Title: Mechanical Engineering; Study programme type: university; Study level: graduate; Academic / professional title: magistar/magistra inženjer/inženjerka strojarstva (magistar/magistra inženjer/inženjerka strojarstva)
• Type of resource: Text
• File origin: Born digital
• Access conditions: Open access
• Created on: 2024-07-13 14:37:43