The study presents a blockchain-based framework to protect AI models for brain-tumor detection from adversarial perturbations in MRI data. It leverages Hyperledger Fabric and Private IPFS to decentralize data storage and access control, ensuring data provenance and tamper resistance. By storing MRI scans on a permissioned blockchain network, the framework aims to prevent unauthorized access and deliberate tampering that could undermine diagnostic accuracy. Experimental evaluations using Hyperledger Caliper demonstrate decentralized cryptographic guarantees of image integrity across healthcare and AI ecosystems.
The dataset used for validation is publicly available in the Br35H: Brain Tumor Detection 2020 repository on Kaggle. This blockchain-driven approach could bolster trust in AI-assisted diagnoses among clinicians and patients. This article outlines a blockchain-based framework designed to shield AI models for brain-tumor detection from adversarial perturbations in MRI data. By leveraging Hyperledger Fabric and Private IPFS, the system decentralizes data storage and access control to strengthen provenance and tamper resistance.
Storing MRI scans on a permissioned blockchain network aims to prevent unauthorized access and deliberate tampering that could undermine diagnostic accuracy. The framework provides cryptographic guarantees of data integrity across healthcare and AI ecosystems, as demonstrated by Hyperledger Caliper evaluations. The validation dataset used is Br35H: Brain Tumor Detection 2020 from Kaggle, illustrating how a blockchain-driven approach could bolster trust in AI-assisted diagnoses among clinicians and patients.
Leave a Reply