@article{ART003370687},
author={Cheolhun Hwang and Geon-Yun Shin},
title={ROB Segmentation-Based Selective Encryption with Random Pattern Fragmentation for Spine X-ray Images},
journal={Journal of The Korea Society of Computer and Information},
issn={1598-849X},
year={2026},
volume={31},
number={8},
pages={73-85}
TY - JOUR
AU - Cheolhun Hwang
AU - Geon-Yun Shin
TI - ROB Segmentation-Based Selective Encryption with Random Pattern Fragmentation for Spine X-ray Images
JO - Journal of The Korea Society of Computer and Information
PY - 2026
VL - 31
IS - 8
PB - The Korean Society Of Computer And Information
SP - 73
EP - 85
SN - 1598-849X
AB - This study proposes a region-of-background (ROB)-oriented selective encryption framework for protecting the PixelData of de-identified Digital Imaging and Communications in Medicine (DICOM) spine X-ray images. A convolutional neural network (CNN) detects the ROB, and its complementary region is defined as the region of interest (ROI). AES-256-GCM is applied to the ROI, AES-128-GCM to the ROB, and random pattern fragmentation (RPF) with 64×64 blocks is used to obscure regional boundaries. The framework was evaluated using 1,200 cervical, lumbar anteroposterior, and lumbar lateral images from the Cervical Spine X-ray Atlas and BUU-LSPINE datasets. The ROB segmentation model achieved a Dice similarity coefficient of 95.67(±0.43)% and an ROI-to-ROB error rate of 1.17(±0.25)%. The encrypted images showed an NPCR of approximately 99.61%, entropy values close to 8.0, and mean absolute adjacent-pixel correlations of 0.00245–0.00253. The mean processing time was 36.27 ms/MP, and all reconstructed images achieved a 100% SHA-256 hash match. The results suggest that the proposed framework may reduce computational burden while maintaining statistical security indicators comparable to previous methods. This study was limited to de-identified two-dimensional spine X-ray images and operational ROB reference masks; DICOM header de-identification and practical key-management implementation were beyond its scope.
KW - Medical image encryption;X-ray encryption;Selective encryption;Random Pattern Fragmentation;Deep learning segmentation
DO -
UR -
ER -
Cheolhun Hwang and Geon-Yun Shin. (2026). ROB Segmentation-Based Selective Encryption with Random Pattern Fragmentation for Spine X-ray Images. Journal of The Korea Society of Computer and Information, 31(8), 73-85.
Cheolhun Hwang and Geon-Yun Shin. 2026, "ROB Segmentation-Based Selective Encryption with Random Pattern Fragmentation for Spine X-ray Images", Journal of The Korea Society of Computer and Information, vol.31, no.8 pp.73-85.
Cheolhun Hwang, Geon-Yun Shin "ROB Segmentation-Based Selective Encryption with Random Pattern Fragmentation for Spine X-ray Images" Journal of The Korea Society of Computer and Information 31.8 pp.73-85 (2026) : 73.
Cheolhun Hwang, Geon-Yun Shin. ROB Segmentation-Based Selective Encryption with Random Pattern Fragmentation for Spine X-ray Images. 2026; 31(8), 73-85.
Cheolhun Hwang and Geon-Yun Shin. "ROB Segmentation-Based Selective Encryption with Random Pattern Fragmentation for Spine X-ray Images" Journal of The Korea Society of Computer and Information 31, no.8 (2026) : 73-85.
Cheolhun Hwang; Geon-Yun Shin. ROB Segmentation-Based Selective Encryption with Random Pattern Fragmentation for Spine X-ray Images. Journal of The Korea Society of Computer and Information, 31(8), 73-85.
Cheolhun Hwang; Geon-Yun Shin. ROB Segmentation-Based Selective Encryption with Random Pattern Fragmentation for Spine X-ray Images. Journal of The Korea Society of Computer and Information. 2026; 31(8) 73-85.
Cheolhun Hwang, Geon-Yun Shin. ROB Segmentation-Based Selective Encryption with Random Pattern Fragmentation for Spine X-ray Images. 2026; 31(8), 73-85.
Cheolhun Hwang and Geon-Yun Shin. "ROB Segmentation-Based Selective Encryption with Random Pattern Fragmentation for Spine X-ray Images" Journal of The Korea Society of Computer and Information 31, no.8 (2026) : 73-85.
