A Novel Visual Cryptography Framework for Secure Blood Report Transmission and Analysis
Anant Manish Singh, Divyanshu Brijendra Singh, Aditya Ratnesh Pandey, Maroof Rehan Siddiqui, Shifa Siraj Khan, Sanika Satish Lad · International Journal of Research Publication and Reviews · 2025
Blood reports constitute critical confidential patient data requiring robust security during storage and transmission.This paper proposes an integration of visual cryptography and advanced image-based encryption tailored for digitized blood reports.The framework leverages an optimized expansion-free halftone-based visual cryptographic scheme to split blood report images into meaningful shares, distributed across multiple storage nodes.A transfer-learning-augmented convolutional neural network (CNN) reconstructs and analyses decrypted reports for diagnostic metrics.Experiments utilize the publicly available BCCD (Blood Cell Count and Detection) dataset (364 images, 640×480 px), extended with 874 augmented samples for robustness.Metrics Peak Signal-to-Noise Ratio (PSNR), Structural Similarity Index (SSIM) and classification accuracy demonstrate superior performance over state-of-the-art medical image VC systems, achieving PSNR of 52.7 dB, SSIM of 0.998 and diagnostic classification accuracy of 95.4%.Comparative analysis with recent VC methods highlights our framework's 15% reduction in share-generation time and 8% higher diagnostic accuracy [2][3][4] .The proposed approach ensures data confidentiality, integrity and availability, mitigating limitations of pixel expansion and contrast loss in traditional VC.