kth.sePublications
Change search
CiteExportLink to record
Permanent link

Direct link
Cite
Citation style
  • apa
  • ieee
  • modern-language-association-8th-edition
  • vancouver
  • Other style
More styles
Language
  • de-DE
  • en-GB
  • en-US
  • fi-FI
  • nn-NO
  • nn-NB
  • sv-SE
  • Other locale
More languages
Output format
  • html
  • text
  • asciidoc
  • rtf
Comparing Convolutional Neural Networks to traditional methods and the human eye for copy-move forgery detection
KTH, School of Electrical Engineering and Computer Science (EECS), Computer Science.
KTH, School of Electrical Engineering and Computer Science (EECS), Computer Science.
2022 (English)Independent thesis Basic level (degree of Bachelor), 10 credits / 15 HE creditsStudent thesisAlternative title
Konvolutionella Neurala Nätverk jämfört med traditionella metoder och det mänskliga ögat för upptäckandet av copy-move förfalskningar (Swedish)
Abstract [en]

A common image forgery technique is called copy-move forgery, in which a part of an image has been copied and placed elsewhere to cover some other part of the image. This study compares implementations of both traditional and deep learning methods for copy-move forgery detection, as well as human ability to do the same. Two Convolutional Neural Networks, one utilizing transfer learning and the other a custom architecture with a so-called constrained layer, are implemented, as well as two traditional methods based on SIFT and BRISK keypoint comparison respectively. The results show that both of the implemented neural networks slightly outperform the traditional methods in terms of accuracy, with a considerably lower computational cost. Additionally, a survey is conducted (n=19) which shows that the neural networks are able to outperform humans to an extent when given the task of classifying images as either forged or authentic. This implies that there is a real practical use for these models, in particular the model utilizing the constrained layer, outperforming the transfer learning model in terms of both accuracy and runtime.

Abstract [sv]

En vanlig förfalskningsteknik kallas copy-move-forgery, där en del av en bild har kopierats och placerats någon annanstans för att täcka någon annan del av bilden. Denna studie jämför implementeringar av både traditionella- och djupinlärningsbaserade metoder för förfalskningsdetektering, såväl som mänsklig förmåga att göra detsamma. Två Konvolutionella Neurala Nätverk, ett som använder transfer learning och det andra en anpassad arkitektur med ett så kallat constrained layer, implementeras, samt två traditionella metoder baserade på jämförelse av SIFT respektive BRISK keypoints. Resultaten visar att båda implementerade neurala nätverk presterar något bättre än de traditionella metoderna vad gäller noggrannhet, med en betydligt lägre beräkningskostnad. Dessutom genomförs en undersökning (n=19) som visar att de neurala nätverken till viss del presterar bättre än människor när de får uppgiften att klassificera bilder som antingen förfalskade eller autentiska. Detta antyder att det finns en verklig praktisk användning av dessa modeller, i synnerhet modellen som använder ett constrained layer, som överträffar transfer learning modellen när det gäller både noggrannhet och exekveringstid.

Place, publisher, year, edition, pages
2022. , p. 35
Series
TRITA-EECS-EX ; 2022:483
National Category
Computer and Information Sciences
Identifiers
URN: urn:nbn:se:kth:diva-319905OAI: oai:DiVA.org:kth-319905DiVA, id: diva2:1702458
Subject / course
Computer Science
Educational program
Master of Science in Engineering - Computer Science and Technology
Supervisors
Examiners
Available from: 2022-10-11 Created: 2022-10-11 Last updated: 2022-10-11Bibliographically approved

Open Access in DiVA

fulltext(2738 kB)222 downloads
File information
File name FULLTEXT01.pdfFile size 2738 kBChecksum SHA-512
ef645fae90b3726dce353046f023bcf4742814a127b9ac544be891aef6a9c4e764cf486dd61c7193121befe8da076e7c902a4b44cf8b1803875c795fc8231897
Type fulltextMimetype application/pdf

By organisation
Computer Science
Computer and Information Sciences

Search outside of DiVA

GoogleGoogle Scholar
Total: 222 downloads
The number of downloads is the sum of all downloads of full texts. It may include eg previous versions that are now no longer available

urn-nbn

Altmetric score

urn-nbn
Total: 385 hits
CiteExportLink to record
Permanent link

Direct link
Cite
Citation style
  • apa
  • ieee
  • modern-language-association-8th-edition
  • vancouver
  • Other style
More styles
Language
  • de-DE
  • en-GB
  • en-US
  • fi-FI
  • nn-NO
  • nn-NB
  • sv-SE
  • Other locale
More languages
Output format
  • html
  • text
  • asciidoc
  • rtf