Image blind detection based on LBP residue classes and color regions

Area of Science:

• Computer Vision
• Digital Image Processing
• Forensic Science

Background:

• Image forgery detection is crucial for verifying digital authenticity.
• Existing block-based methods offer high accuracy but suffer from heavy computational load due to exhaustive searches.
• Novel approaches are needed to enhance efficiency without compromising detection rates.

Purpose of the Study:

• To develop an efficient and accurate image forgery detection method.
• To reduce the computational complexity of existing techniques.
• To improve the speed of detecting tampered regions in digital images.

Main Methods:

• The proposed method utilizes local binary pattern residue classes and color regions for image analysis.
• Images are divided into overlapped blocks, and similar blocks are identified based on shared characteristics.
• Suspicious regions are analyzed, and morphological operations are used to refine the detection of tampered areas.

Main Results:

• The method demonstrated improved detection accuracy compared to traditional approaches.
• Execution time was significantly reduced due to a narrowed search range for similar blocks.
• The technique proved effective under various challenging image conditions.

Conclusions:

• The developed forgery detection method offers a balance of high accuracy and improved efficiency.
• The approach effectively reduces computational load by optimizing the search for forged regions.
• This technique presents a promising advancement in digital image forensics.