Structure-Aware Hierarchical Sha-256 Hashing for Dicom Integrity Verification in Consortium Blockchain Systems

Abstract

Integrity verification of DICOM-based medical data is a critical requirement in distributed healthcare systems, where conventional hashing methods operate at the file level and do not account for the internal structure of medical images. Existing approaches treat DICOM files as undifferentiated byte streams, limiting sensitivity to localized modifications in metadata and pixel-level content. This study addresses this limitation by evaluating a structure-aware hierarchical hashing scheme that incorporates intrinsic DICOM attributes into the hashing process. The objective is to assess whether integrating metadata features, pixel-based descriptors, and entropy measures within a hierarchical SHA-256 framework influences the statistical characteristics of hash outputs. An experimental comparative design is implemented using 427 DICOM files. The proposed method applies a multi-component HMAC-SHA256-based hashing pipeline and is evaluated against standard SHA-256, SHA-3, and locality-sensitive hashing (LSH). Evaluation metrics include entropy (bounded between 0 and 8), bit distribution, chi-square statistics, Kolmogorov–Smirnov test results, avalanche effect, and execution time. The results indicate that all methods maintain statistically valid output distributions, while the proposed approach exhibits greater variability in entropy values within the valid range, consistent bit balance, lower deviation in distributional tests relative to LSH, and stable avalanche behavior across structured input variations. Execution time increases moderately due to multi-stage processing. These findings show that hierarchical decomposition produces enhanced sensitivity to input structure without altering deterministic hashing properties. The study contributes a domain-aware hierarchical hashing framework that integrates DICOM structural components into hash generation. This contribution demonstrates that structure-aware hashing enables more sensitive detection of input-level variations while preserving statistical correctness, providing a methodologically grounded approach for integrity verification in medical data systems without extending claims to cryptographic superiority.