Automatic Selective Encryption of DICOM Images

• Published in: Applied sciences 2023-04, Vol.13 (8), p.4779
• Main Authors: Natsheh, Qamar, Sălăgean, Ana, Zhou, Diwei, Edirisinghe, Eran
• Format: Article
• Language: English
• Subjects: Algorithms; Compression; Compression ratio; Confidentiality; Data encryption; LZ4 compression; Mathematical models; medical image encryption; Medical imaging; Medical imaging equipment; multi-frame DICOM object; Multimedia; Optimization; Privacy; Privacy, Right of; region of interest; Safety and security measures; Security management; segmentation; selective encryption
• ISSN: 2076-3417; 2076-3417
• DOI: 10.3390/app13084779

Securing DICOM images is essential to protect the privacy of patients, especially in the era of telemedicine and eHealth/mHealth. This increases the demand for rapid security. Nevertheless, a limited amount of research work has been conducted to ensure the security of DICOM images while minimizing the processing time. Hence, this paper introduces a selective encryption approach to reduce the processing time and sustain the robustness of security. The proposed approach selects regions within medical images automatically in the spatial domain using the pixel thresholding segmentation technique, then compresses and encrypts them using different encryption algorithms based on their importance. An adaptive two-region encryption approach is applied to single and multi-frame DICOM images, where the Region of Background (ROB) is encrypted using a light encryption algorithm, while the Region of Interest (ROI) is encrypted using a sophisticated encryption algorithm. For multi-frame DICOM images (Approach I), additional time-saving has been achieved by almost 10,000 times faster than the Naïve encryption approach, and 100 times better compression ratio, using one segmentation map based on a pre-defined reference frame for all the DICOM frames. For single-frame DICOM image (Approach II), a multi-region selective encryption approach is proposed, where the ROI is further split into three regions based on potential security threats, using a mathematical model that guarantees shorter encryption time in comparison with the Naive and the two-region encryption approaches, with almost 47% and 14% saving times, respectively. Based on the estimated processing time, Approach I outperformed Approach II noticeably. Further, cryptanalysis metrics are utilized to evaluate the proposed approaches, which indicate good robustness against a wide variety of attacks.