Debashis Das

2papers

2 Papers

CRFeb 4, 2023
Use of Federated Learning and Blockchain towards Securing Financial Services

Pushpita Chatterjee, Debashis Das, Danda B Rawat

In recent days, the proliferation of several existing and new cyber-attacks pose an axiomatic threat to the stability of financial services. It is hard to predict the nature of attacks that can trigger a serious financial crisis. The unprecedented digital transformation to financial services has been accelerated during the COVID-19 pandemic and it is still ongoing. Attackers are taking advantage of this transformation and pose a new global threat to financial stability and integrity. Many large organizations are switching from centralized finance (CeFi) to decentralized finance (DeFi) because decentralized finance has many advantages. Blockchain can bring big and far-reaching effects on the trustworthiness, safety, accessibility, cost-effectiveness, and openness of the financial sector. The present paper gives an in-depth look at how blockchain and federated learning (FL) are used in financial services. It starts with an overview of recent developments in both use cases. This paper explores and discusses existing financial service vulnerabilities, potential threats, and consequent risks. So, we explain the problems that can be fixed in financial services and how blockchain and FL could help solve them. These problems include data protection, storage optimization, and making more money in financial services. We looked at many blockchain-enabled FL methods and came up with some possible solutions that could be used in financial services to solve several challenges like cost-effectiveness, automation, and security control. Finally, we point out some future directions at the end of this study.

CRMay 30, 2012
Steganography Using Adaptive Pixel Value Differencing(APVD) of Gray Images Through Exclusion of Overflow/Underflow

J. K. Mandal, Debashis Das

In a gray scale image the pixel value ranges from 0 to 255. But when we use pixel-value differencing (pvd) method as image steganographic scheme, the pixel values in the stego-image may exceed gray scale range. An adaptive steganography based on modified pixel-value differencing through management of pixel values within the range of gray scale has been proposed in this paper. PVD method is used and check whether the pixel value exceeds the range on embedding. Positions where the pixel exceeds boundary has been marked and a delicate handle is used to keep the value within the range. From the experimental it is seen that the results obtained in proposed method provides with identical payload and visual fidelity of stego-image compared to the pvd method.