الفهرس 
Literature Review of Image Security ApproachesOnly 14 pages are availabe for public view
 |  | from | 221 |  |  |
| | | from | 221 | | |
Abstract
In today’s world, health information systems form a critical part of information
technology infrastructure in any country. Medical diagnostics through
telecommunication is recommended to play a major role in people life. With today’s
advancement in health systems, remote access to hospital data and cloud storage of
healthcare data and imaging data is a key to effective healthcare delivery. The critical
nature of health services and opportunities provided by software applications makes it
possible for healthcare delivery to be introduced with more efficiency and less time. In
recent times, securing medical images in telemedicine applications has become a
challenging issue.
In the first part of this thesis, medical image cryptography is utilized in various
healthcare and Internet of Medical Things (IoMT) applications. The proposed
encryption method for medical images depends on a 3D chaotic-based non-linear
ciphering process, which is used for pixel value diffusion and position permutation.
In the second part of the thesis, a DNA encryption algorithm is presented for Internet
of Things (IoT) applications. Several telemedicine IoT applications are expected to be
adopted in the medical sector in the incoming several years. This algorithm is an
efficient cryptosystem framework for medical image encryption based on chaos and
DNA encoding. To produce all characteristics required from this encryption algorithm,
the Piecewise Linear Chaotic Map (PWLCM), Logistic Map, and DNA encoding
functions are employed.
In the third part of this thesis, a robust multi-level security framework based on
fusion, encryption, and watermarking processes is presented for efficient transmission
of color medical images. The proposed watermarking technique depends on the mixture
of Discrete Cosine Transform (DCT), Lifting Wavelet Transform (LWT), and Singular
Value Decomposition (SVD).
The fourth part of this thesis presents a medical image cryptosystem, which uses a
Stacked Auto-Encoder (SAE) network to produce two sets of chaotic random matrices.The first set is utilized to generate a complete shuffling matrix that will change the pixel
locations on the input digital image. The second set generates an independent series of
sequences that are employed to enlarge the difference between the permutated
encrypted medical images and the original images. The proposed cryptosystem is robust
due to the benefits of the parallel SAE computations, which significantly reduces the
runtime complexity.