Medical image watermarking with ANN in wavelet domain
P. V. V. Kishore, K. Sai Prajwal, M. Kamal Mohan, S. Koteswarao · 2015
Medical images hold health information about a patient. Due to their inability to show information clearly and lack of expert doctors, motivate patients to send their imaging reports using unsecured Internet. The objective is to provide security to medical images of patients passing through unsecured networks through watermarking methods. The watermarking will most likely will alter pixels in the medical image reducing object composition of the medical image. Medical image quality affects the health of a patient. To avoid such quality issues during watermarking, the algorithm uses artificial neural network (ANN). ANN remembers and rebuilds the degraded pixel information by its location during watermarking and transmission. Embedding and extraction of watermark uses wavelet domain. Experimentation performed on three types of medical images Ultrasound imagery (US), Magnetic Resonance Imaging (MRI) and Computer Tomography (CT) of test subjects brought from hospitals in Vijayawada in India. The psnr and normalized cross correlation coefficient (ncc) measures the visual quality and quantity of the watermarking procedure with ANN. The DWT-ANN based method preserves the quality of the medical images.