Enhancing Blockchain Security Against Data Tampering: Leveraging Hybrid Model in Multimedia Forensics and Multi-Party Computation for Supply Chain Data Protection

Over the past few years, there has been a notable surge in the integration of Blockchain technology into supply chain management systems. This integration holds the promise of enhanced transparency, security, and efficiency in monitoring the movement of goods and services. This study presents a novel approach aimed at fortifying privacy and accuracy within blockchain-based supply chain management systems. The methodology integrates Long Short-Term Memory (LSTM) and Gated Recurrent Unit (GRU) units with secure multi-party computation (MPC) and differential privacy techniques as a hybrid model. The objective is to safeguard the confidentiality of transaction data while enabling precise detection of media tampering. Performance evaluation revolves around three key aspects: accuracy, privacy preservation, and computational efficiency. In terms of accuracy assessment, the proposed hybrid approach is benchmarked against traditional machine learning algorithms including Support Vector Machines (SVM), k-Nearest Neighbors (KNN), and Random Forest. Results indicate superior performance, with the proposed hybrid method achieving an accuracy of 0.95, outperforming conventional algorithms. Precision, recall, and F1-score metrics further confirm the effectiveness of the approach in accurately identifying media tampering instances. Privacy preservation capabilities are evaluated through differential privacy techniques, revealing the method's ability to inject controlled noise into the data to protect individual privacy. Results demonstrate varying levels of privacy preservation across different settings, highlighting the trade-off between privacy and data utility. Computational efficiency is also scrutinized, considering the additional overhead introduced by privacy preservation mechanisms and secure MPC protocols. While there is a slight increase in computational time, the proposed approach maintains reasonable training and inference times, ensuring practical applicability in real-world scenarios.