Zero-bit fragile watermarking for medical image tamper detection and recovery using RS code and lifting wavelet transform

ABSTRACT

Medical images are habitually composed of Regions of Interest (ROI) and Region of Non-interest (RONI). ROI means the area containing the crucial information for diagnosis. To achieve the fidelity and integrity of the ROI, we propose a novel watermarking approach based on two great concepts zero-watermarking and Error Correcting Code (ECC). Tamper detection and recovery information are stored in RONI which accepts some visual quality degradation. Experimental results prove that our proposed approach produces a good performance in terms of imperceptibility and the ability to resist several hard attacks. The proposed scheme can recover 100% of the ROI until 15% of distortion is caused by salt and pepper noise. For multiplicative and Gaussian noise, the scheme has also a good capability of recovery (between 70% and 100%). With close to 20% of distortion, the system can recover 50% of the ROI. The proposed method has also a good capability of recovery against cropping and image processing operations.

KEYWORDS:

• Zero watermarking
• tamper detection
• recovery
• Error Correcting Code (ECC)
• Reed Solomon corrector code (RS)
• Lifting Wavelet Transform (LWT)
• medical image
• region of interest

Disclosure statement

No potential conflict of interest was reported by the author(s).

Additional information

Notes on contributors

Nour El Houda Golea

Nour El-Houda Golea obtained her engineer's degree in computer science in 2007 from the University of Batna, Algeria. She received the Master and the Ph.D degrees from Batna University, in 2010 and 2018, respectively. She is currently a senior lecturer in Computer Science Department, Faculty of Mathematics and Computer Science, Batna2 University, Algeria. Her current research interests include medical imaging, digital watermarking, image processing and digital content protection.

Kamal Eddine Melkemi

Kamal Eddine Melkemi received his engineer's degree (bacaloréat+5 years) and his Master degree, both in computer science institute from Annaba University and Batna University, Algeria, in 1990 and 1993, respectively. He received his doctorate (Ph.D. degree) in computer science from the Constantine University in 2006. His research dissertation is about evolutionary algorithms based on MRF for image segmentation. He is a Member of IEEE QCIT and a permanent reviewer of the journal CAIE Elsevier. His current research interests include pattern recognition, artificial intelligence, machine and deep learning, robotics.

Ali Behloul

Ali Behloul is a full Professor of Computer Science at Batna 2 University, Algeria. In 2007, he received his PhD in Computer Science from the university of Orsay at Paris, France. His main research interests include Information Retrieval, Pattern Recognition, Image Processing, Artificial Intelligence and Image Watermarking. He is a member of the LaSTIC Laboratory.