LPCLNet: Leveraging local pixel-wise contrastive learning for image tampering localization

Puntos clave

• Image tampering localization improves significantly using the proposed algorithm, enhancing detection accuracy.
• Key metrics show an increase in localization performance, with a notable improvement over existing methods.
• Assessments using the LPCLNet approach highlight the effectiveness of local pixel-wise contrastive learning in image detection.
• Results emphasize the need for advanced techniques in digital forensics, particularly for real-world applications.