Date of Award

Summer 7-29-2025

Document Type

Thesis

Degree Name

Master of Science (MS)

Department/Program

Digital Forensics and Cybersecurity

Language

English

First Advisor or Mentor

Shweta Jain

Second Reader

Hunter Johnson

Third Advisor

Douglas Salane

Abstract

Perceptual hashing algorithms are algorithms that generate content-based image hashes by extracting perceptual features from the images. Unlike cryptographic hashes, which exhibit significant changes with even slight input alterations, perceptual hashes do not change when modifications like compression, color correction and brightness are applied to the images. These hashes are designed to remain similar for inputs that are visually or perceptually alike, which has led to their widespread application in detecting duplicate images, finding similar images for reverse image search and to detecting inappropriate content of Child sexual abuse (CSAM) images by comparing image hashes with dataset of known perceptual hashes. This study focuses on evaluating the robustness of NeuralHash against adversarial attacks including Hash collision attack, standard evasion attack, edges only attack and few pixels attacks. Attacks have been conducted on CASIA dataset with the aim of measuring how effectively NeuralHash resists these adversarial modifications while maintain visual consistency in the image data.

Share

COinS
 
 

To view the content in your browser, please download Adobe Reader or, alternately,
you may Download the file to your hard drive.

NOTE: The latest versions of Adobe Reader do not support viewing PDF files within Firefox on Mac OS and if you are using a modern (Intel) Mac, there is no official plugin for viewing PDF files within the browser window.