Title

Automated Analysis of Encrypted and Obfuscated Data based on Deep Neural Networks

Alternative Title

딥러닝 기반의 암호화 및 난독화 된 데이터 자동 분석

DGIST Authors

Ongee Jeong ; Inkyu Moon ; Goo-Rak Kwon

Advisor

문인규

Co-Advisor(s)

Goo-Rak Kwon

Issued Date

2025

Awarded Date

2025-02-01

Citation

Ongee Jeong. (2025). Automated Analysis of Encrypted and Obfuscated Data based on Deep Neural Networks. doi: 10.22677/THESIS.200000841197

Type

Thesis

Description

Deep Learning, Data Analysis, Cryptanalysis, Privacy-Preserving

Table Of Contents

Ⅰ. INTRODUCTION 1
1.1. Motivations and Objectives 1
1.2. Overview 5
1.3. Contributions and Outline 7

Ⅱ. DEEP LEARNING-BASED ENCRYPTED DATA ANALYSIS 9
2.1. Deep Learning-based Cryptanalysis on Optical Cryptographic Algorithm 9
2.1.1. Methodology 9
2.1.2. Experiments 16
2.2. Deep Learning-based Cryptanalysis on Block Ciphers 23
2.2.1. Methodology 23
2.2.2. Experiments 36
2.3. Deep Learning-based Cryptanalysis on Public-Key Cryptography 49
2.3.1. Methodology 49
2.3.2. Experiments 52

Ⅲ. DEEP LEARNING-BASED OBFUSCATED DATA ANALYSIS 64
3.1. Methodology 64
3.1.1. Poisson-Multinomial Distribution-based Photon Counting Imaging (PMD-PCI) 64
3.1.2. Deep Learning-based Privacy-Preserving Image Classification Scheme 65
3.2. Experiments 70
3.2.1. Dataset 70
3.2.2. Implementation Details 70
3.2.3. Evaluation Metric 71
3.2.4. Results 72

Ⅳ. CONCLUSION AND FUTURE WORK 81
4.1. Summary and Discussion 81

References 85

요 약 문 91

URI

http://hdl.handle.net/20.500.11750/57969
http://dgist.dcollection.net/common/orgView/200000841197

DOI

10.22677/THESIS.200000841197

Degree

Doctor

Department

Department of Robotics and Mechatronics Engineering

Publisher

DGIST

Show Full Item Record