A novel forged image detection method using the characteristics of interpolation

Area of Science:

• Digital Forensics
• Image Processing
• Computer Vision

Background:

• The proliferation of digital image editing tools has led to widespread exposure to forged images.
• Dissemination through various media channels means unverified images are often accepted as truth, causing societal issues.
• Existing methods for detecting image forgeries require improvement to address sophisticated manipulation techniques.

Purpose of the Study:

• To develop a reliable method for identifying the veracity of digital images.
• To propose a novel re-interpolation algorithm for forgery detection.
• To enhance the accuracy and efficiency of digital image forensics.

Main Methods:

• A re-interpolation algorithm was developed, applying identical arbitrary values to both interpolated and un-interpolated image regions.
• The study analyzed the discrete Fourier transform (DFT) characteristics of these regions to identify discrepancies.
• A detection map was generated by calculating the subtraction value between DFT characteristics of the two regions.

Main Results:

• The proposed method successfully distinguished between original and forged image regions.
• The discrete Fourier transform (DFT) analysis revealed unique characteristics indicative of image manipulation.
• The generated detection map effectively highlighted areas of forgery within digital images.

Conclusions:

• The re-interpolation algorithm combined with discrete Fourier transform (DFT) analysis offers a robust approach to digital image forgery detection.
• This method can help mitigate the spread of misinformation from manipulated images.
• Further research can explore real-time application and integration into existing digital forensic tools.