diplomsko delo
Žiga Šolar (Author), Borut Batagelj (Mentor)

Abstract

Diplomska naloga zajema opis problema, razvoj in testiranje rešitve konvolucijske nevronske mreže za enoslikovno super resolucijo v digitalni forenziki. Premajhna ločljivost slik in ostalih slikovnih digitalnih dokazov je pogosti pojav na področju digitalne forenzike. V nalogi so predstavljene obstoječe rešitve metod enoslikovne super resolucije z uporabo globokega strojnega učenja. Naloga se nato osredotoči na implementacijo rešitve izboljšave nizkoločljivih slik človeških obrazov s pomočjo konvolucijske nevronske mreže, ter primerjavo te rešitve z obstoječimi.

Keywords

super resolucija;CNN;nevronske mreže;samokodirnik;prepoznavanje oseb;identifikacija;verifikacija;računalništvo in informatika;visokošolski strokovni študij;diplomske naloge;

Data

Language: Slovenian
Year of publishing:
Typology: 2.11 - Undergraduate Thesis
Organization: UL FRI - Faculty of Computer and Information Science
Publisher: [Ž. Šolar]
UDC: 004.85:343.98(043.2)
COBISS: 75923971 Link will open in a new window
Views: 282
Downloads: 50
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: Super resolution of face images in digital forensics
Secondary abstract: The thesis includes the problem description, development and testing of a convolutional neural network for single image super-resolution in digital forensics. Insufficient resolution of images and other image digital evidence is a common occurrence in the field of digital forensics. This paper presents previously developed solutions for single image super-resolution using Deep Learning based on neural networks. It then describes the implementation of a solution for quality enhancement in low-resolution images of human faces using convolutional neural network and the comparison of the solution results with existing solutions.
Secondary keywords: super-resolution;machine learning;CNN;neural networks;autoencoder;person identification;identification;verification;computer science;computer and information science;diploma;Strojno učenje;Umetna inteligenca;Računalniška forenzika;Računalništvo;Univerzitetna in visokošolska dela;
Type (COBISS): Bachelor thesis/paper
Study programme: 1000470
Embargo end date (OpenAIRE): 1970-01-01
Thesis comment: Univ. v Ljubljani, Fak. za računalništvo in informatiko
Pages: 39 str.
ID: 13324748