• Mr. M. Visweswaran
• Dr. T. Geetha

Digital images face increasing threats from tampering attacks, compromising their authenticity and security, especially on online platforms. Image tampering involves unauthorized modifications that distort content, making it challenging to restore the original integrity using conventional methods. Existing techniques often struggle with compressed or low-resolution images and lack self-recovery capabilities, emphasizing the need for advanced solutions. This project introduces the Photo Forgery Detection Network (PFDNet), a deep learning-based framework designed to detect tampering and enable lossless recovery. PFDNet incorporates a Cyber Vaccinator module, the original image and its edge map are transformed into an immunized version, ensuring consistency with the original content and Forgery Detector module using the Invertible Neural Network for enhanced tamper resistance and self-recovery. In Invertible Neural Network of the forward pass, when an attacked image is received, a localizer predicts a tamper mask to identify altered regions. In the backward pass, the generator converts hidden perturbations into recoverable information, enabling the restoration of the original image along with its edge map. To ensure lossless recovery, Run-Length Encoding (RLE) is employed as a final step to compare the original image and the recovered image. This comparison validates the restoration process, ensuring high fidelity and accuracy. Experimental results demonstrate PFDNet’s ability to accurately localize tampered areas and achieve high-quality image recovery, providing a robust solution for combating image forgery and maintaining digital media integrity.

Digital Forgery, Cyber Vaccinator, PFDNet, Invertible Neural Network, Image Recovery, Run-Length Encoding.

"PFDNET: A DEEP LEARNING APPROACH FOR ROBUST SHARED PHOTO AUTHENTICATION AND TAMPER RECOVERY", International Journal of Emerging Technologies and Innovative Research (www.jetir.org), ISSN:2349-5162, Vol.12, Issue 6, page no.g493-g497, June-2025, Available :http://www.jetir.org/papers/JETIR2506659.pdf

"PFDNET: A DEEP LEARNING APPROACH FOR ROBUST SHARED PHOTO AUTHENTICATION AND TAMPER RECOVERY", International Journal of Emerging Technologies and Innovative Research (www.jetir.org | UGC and issn Approved), ISSN:2349-5162, Vol.12, Issue 6, page no. ppg493-g497, June-2025, Available at : http://www.jetir.org/papers/JETIR2506659.pdf