SRM University-AP is pleased to announce that Dr V M Manikandan, Assistant Professor in the Department of Computer Science and Engineering, along with his research scholar Shaiju Panchikkil, published a paper titled, “A convolutional neural network model based reversible data hiding scheme in encrypted images with block-wise Arnold transform” in Optik Journal, Elsevier Publications.
The proposed scheme can be used for embedding electronic patient reports (EPR) in the medical image itself while transmitting, and at the receiver side the reports can be extracted along with the lossless recovery of the medical images.
About the Research:
Data hiding or information hiding is a well-explored way to secure some secret data by concealing it in a digital cover medium. The reversible data hiding (RDH) is a recent advancement in the field of data hiding in which the cover medium can be restored during the extraction of hidden messages at the receiver side. The RDH schemes are widely used in medical image transmission and cloud computing. Recently, research in the field of RDH in encrypted images got much attention to improve the efficiency parameters such as embedding rate and bit error rate without compromising the lossless recovery of the images. In this research paper, we propose a new RDH scheme in encrypted images which utilizes the Arnold scrambling technique for data hiding. A convolutional neural network (CNN) model is trained and used to extract the hidden message along with the recovery of the original image. The experimental study and result analysis of the proposed scheme are carried out on the USC-SIPI image dataset managed by the University of Southern California.
Dr Manikandan collaborated with Prof. Yu-Dong Zhang, Professor in School of Computing and Mathematical Sciences, University of Leicester, UK collaborated for this research work. In the future, Dr Manikandan’s research will be more focussed on coming up with new reversible data hiding schemes in encrypted images with better embedding rate and bit error.