Dr.  Mansi Siddharth  Subhedar
Pillai HOC College of Engineering and Technology,  India

Title:

Abstract:

The talk will focus on transform domain image steganography algorithms using advanced image transforms like contourlet transform, curvelet transform, and ridgelet transform that offer better directionality, and anisotropy and are suitable to model images rich in directional and edge details. Analysis and simulation-based experimental results demonstrate that the invisibility of a stego image with secret data in it is excellent and the trade-off between payload capacity and robustness can be adjusted according to the need of the application. To further improve the results, the role of matrix decomposition techniques like singular value decomposition and QR factorization was investigated. Extensive experimental investigations confirmed that the proposed work outperformed existing work.

Cover selection helps steganalyser to misclassify stego images as clean images and thus improves security. In literature, sparse information is available on this resource and hence this issue was focused. Soft computing tools like fuzzy logic and neural network were employed to make appropriate choices of the cover image from the image database to reduce the risk of detection. These cover selection methods were based on image complexity and heterogeneity and can be used in the future to identify the set of images to be used for any application. Designed cover selection algorithms were further integrated with content-adaptive embedding techniques to develop image steganography methods. Detection accuracy offered by proposed steganography algorithms was computed using various steganalysis schemes available in the literature. Poor detection accuracy in the presence of a step analyzer; better is the steganography scheme. Owing to this, numerous experiments were carried out and detection accuracy was found to be less as compared to existing work while maintaining a large payload. Experimental results match or outperform the current JPEG domain and other transform domain schemes in all aspects.

Biography:

Dr. Mansi Subhedar completed BE (ETC), ME (Electronics Engineering), and Ph.D. in Electronics Engineering. She has more than 17 years of teaching experience. She is currently working as IQAC Coordinator and Head, at the Department of Electronics and Computer Science at Pillai HOC College of Engineering and Technology, Rasayani, Maharashtra, India. She has published 43 papers in peer-reviewed international journals and conferences. She has received more than 850 citations for her research work to date. She is also a reviewer for international journals (Elsevier, Springer, Taylor, and Francis, etc) and conferences of repute. She was a reviewer and technical committee member of reputed conferences under IEEE, and Springer. She is a life member of ISTE, IETE, IE, and CSI and a Senior Member, of IEEE. 

Her research interests include signal processing, Cyber Security, IoT and Data Science, and next-generation networks.