magistrsko delo
Leon Abraham (Author), Aleš Holobar (Mentor), Aljaž Frančič (Co-mentor)

Abstract

Magistrsko delo opisuje celotni postopek izdelovanje sistema za iskanje oglasa v toku posnetkov. Vsebuje predstavitev predobdelave oglasa ter posnetkov, iskanje slike v posnetku z lastno implementacijo ter s pomočjo storitve AWS Rekognition, uporabniški vmesnik ter predstavitev rezultatov oziroma testiranje sistema. Aplikacija deluje na podlagi iskanja dveh podobnih slik kjer v primeru podobnosti uporabnik dobi določene meta podatke (čas ujemanja, odstotek ujemanja ter ime datoteke) ter sliko, kdaj se je pojavil oglas v toku posnetkov. Sistem smo ovrednotili z mero F1, kjer smo dobili 100% uspešnost.

Keywords

strojno učenje;zaznava oglasov;storitev WS;tehnologija FFMPEG;magistrske naloge;

Data

Language: Slovenian
Year of publishing:
Typology: 2.09 - Master's Thesis
Organization: UM FERI - Faculty of Electrical Engineering and Computer Science
Publisher: [L. Abraham]
UDC: 004.93:004.85(043.2)
COBISS: 183114755 Link will open in a new window
Views: 249
Downloads: 38
Average score: 0 (0 votes)
Metadata: JSON JSON-RDF JSON-LD TURTLE N-TRIPLES XML RDFA MICRODATA DC-XML DC-RDF RDF

Other data

Secondary language: English
Secondary title: Development of ad detection system with machine learning algorithms
Secondary abstract: The master's thesis describes the entire process of creating a system for searching of advertisements in video streams. It includes the presentation of advertisement and video preprocessing, searching for an image within a video using a custom implementation and with the help of AWS Rekognition service, the user interface, and the presentation of results or system testing. The application operates based on searching for two similar images where, in the case of similarity, the user receives certain metadata (time of match in the video stream, percentage of match, and the file name) and the image of the advertisement that appeared in the video stream. The system was evaluated using the F1 measure, where it achieved a 100% success rate.
Secondary keywords: machine learning;ad detection;AWS;FFMPEG;
Type (COBISS): Master's thesis/paper
Thesis comment: Univ. v Mariboru, Fak. za elektrotehniko, računalništvo in informatiko, Računalništvo in informacijske tehnologije
Pages: 1 spletni vir (1 datoteka PDF (X, 61 f.))
ID: 21606495