Please use this identifier to cite or link to this item:
http://hdl.handle.net/123456789/253
Full metadata record
DC Field | Value | Language |
---|---|---|
dc.contributor.author | Biradar, Sumangala | - |
dc.date.accessioned | 2024-01-05T05:31:10Z | - |
dc.date.available | 2024-01-05T05:31:10Z | - |
dc.date.issued | 2022-11 | - |
dc.identifier.uri | http://hdl.handle.net/123456789/253 | - |
dc.description.abstract | The advancement in wireless communication networks like 5G and enhancement in bandwidth, leads to rapid development in telemedicine and e-Health services. These services are increasingly being exploited in geographically expanded areas. Where accessibility of the health services are tremendously diminished or even non-existent. The practitioner or medical experts and patient are not physically in communication and their interactions are medicated through electronic forms in a consolidated remote location. In medical imaging, image types like MRI, CT or ultrasound are used by medical experts for quick and accurate diagnosis of various diseases. The medical images are very important for disease diagnosis, treatments and research. They act as key component of patient records in electronic form. The communication of these medical images is carried out through an insecure opensource network i.e. internet. The digital broadcast of medical image is continuously prone to malwares, cyber criminals and other infringements of security. The communication of medical image across vulnerable channel adds potential risk of undesirable effect and causes important disease related information in the medical image being lost or corrupted. Then, it becomes problematic for medical experts to diagnose the specific disease from corrupted or lost part of medical image. Therefore, security for medical image has become more essential and must fulfil the requirements like integrity, reliability and confidentiality. Several state-of-art encryption techniques are available. Due to intrinsic properties of medical images such as mass data volume, high pixel correlation among adjacent pixels and high redundancy, these encryption algorithms are not appropriate for medical image encryption. Nowadays a main research challenge is about providing the security, integrity and confidentiality for medical images. Deoxyribonucleic Acid (DNA) cryptography is an advanced evolving technology in cryptography. It is depending on DNA operations and DNA sequences. In DNA cryptography, data is hidden in terms of DNA sequences by performing biological process. The uniqueness of DNA is appropriate to offer security for medical images. The main aim of the research study is to develop an efficient and effective system to provide the enhanced higher-level security, integrity and confidentiality for medical images. The xiv system helps for secure transmission of medical images in real-time applications like telemedicine, e-health services etc. The workflow consists of, enhancement in security of medical images, enforcing integrity for tamperproof medical images and confidentiality of patient information to prevent from eavesdropper. Along with improvement of security, reduction in computation time is also essential for transmission of medical image through network in less time. The research work starts with generation of key image using pseudorandom generator. The key image and original medical image are renovated into encoded DNA structures using fixed DNA encoding rules. The pixels of matrices are permuted using Chen’s chaotic sequences and diffused using DNA XOR operation. The diffused encoded DNA structures is renovated into cipher image using fixed DNA decoding rules. The performance parameters number of pixels changing rate (NPCR), unified average change intensity (UACI), Chi-square, correlation coefficient, entropy, mean square error (MSE), and peak signal to noise rate (PSNR) are used for security analysis. The key space is sufficient to endure against brute force attack. The time and space efficiency of this method is burden. To offer superior security for medical image, the original medical image is segmented into ODD and EVEN images based on intensity levels. The fixed DNA encoding rules are applied to construct ODD and EVEN encoded DNA structures. The multiple hyper chaotic sequences are used for permutation process. Encoded DNA structures are diffused by DNA ADD operation. The fixed DNA decoding rule is applied to get the cipher image. The key space is increased due to multiple chaotic sequences. The integrity and confidentiality are not verified. The hash function SHA-256 and SHA-512 generates a hash key for the specific DNA sequence. This hash key is useful to enforce integrity, and specific DNA sequence is useful for confidentiality. For the enhancement of security level of medical image, multilevel security has been developed. The original medical image is bifurcated into two sub images. These sub images are transformed into encoded DNA structures using dynamic DNA encoding rules. The multiple chaotic sequences are utilized for the permutation of encoded DNA structures in permutation process. In diffusion process, DNA ADD operation is applied for the diffusion of structures. The dynamic DNA complementary rules are used to gain a cipher image. The time efficiency of this crypto method is high. xv To reduce the time required for encryption algorithm, selective medical image encryption using dual hyper chaos map, dynamic DNA encoding, DNA XOR operation and dynamic complementary rules are proposed. These techniques help to reduce the time required with compromising in security. To provide significant security with reduced time, discreet Haar wavelet transform is proposed to compress the medical image. The dual hyper maps are employed for permutation process and DNA XOR operation for diffusion purpose. The dynamic complementary rules are used to get a cipher image. This leads to enhancement in security with compromising in image quality. These en/decryption methods are implemented sequentially. In sequential computation, these multilevel approaches take more times to accomplish enhanced security for medical images. For providing the significant security for medical image with less time, the parallel approach is proposed. The original medical image is fragmented into four sub images. These sub images are renovated into DNA sequence matrices using dynamic DNA encoding rules. The pixels of DNA sequence matrices are permuted using four-dimensional multiple chaotic sequences. The DNA XOR operation is applied for diffusion of the DNA sequence matrices. The hash function SHA-512 generates four hash keys, for four different DNA sequences. These keys are used to enforce integrity and four different DNA sequences are used for confidentiality. In parallel approach, multiple threads are created to run encryption process for four sub images parallelly, which reduces the computation time with enhanced security. The performance of the developed cryptosystems are analyzed using security analysis and time taken for execution. The security analysis depends on resistant of proposed cryptosystems against statistical attacks, differential attacks, and exhaustive attacks. The performance parameters namely, correlation coefficient, histogram analysis, NPCR, UACI, key space, and key sensitivity are utilized to prove the resistant against various attacks. The quality of cryptosystems is measured using MSE, PSNR and entropy. The parallel approach for encryption method is better for providing enhanced higher-level security for medical images with reduced time and resistant to crypto attacks. The work carried out is useful for real-time applications like, virtual consultancy, telemedicine and e-health systems | en_US |
dc.language.iso | en_US | en_US |
dc.title | A Study of Multilevel DNA Cryptosystem for Medical Images | en_US |
dc.type | Thesis | en_US |
Appears in Collections: | Ph.D Thesis |
Files in This Item:
File | Description | Size | Format | |
---|---|---|---|---|
THESIS_2BL17PEA03_SUMANGALA_BIRADAR-1.pdf | 13.78 MB | Adobe PDF | View/Open |
Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.