Reversible Protection and Authentication Techniques for Digital Medical Images

Digital medical images can get various informational applications and take the advantages of cyber-net to enhance the core competence of medical treatment and save plenty resources. However, there are a lot of information securities concerns like: stealing, tampering and counterfeiting when the digital medical images are preceded, stored, printed or transmitted on cyber-net. Therefore, applying the information hiding technology such as steganography and digital watermarking techniques for providing security, integrity, non-repudiation and authentication under the specific standard and special demands for those digital imaging with critical information is a significant and urgent research issue. The digital image protection and authentication techniques are the newest technologies of information hiding area in recent years. The main idea of those technologies is using steganography methods to add protective or authenticated information that can，t be found by human eyes into digital images. With those adding information we can guarantee the reliability of the protective image. The digital image technology of protection and authentication can satisfy the need of high reliability requirements of digital media such as the work of art, military, and medical domain. The protection and authentication methods that applied to medical-images are the last defense line. When medical research or examination, the malicious change or the crucial difference which comes with embedding digital watermarking to original image could let doctors make the misdiagnosis and lead patient facing the risk of medical malpractice. The steganography methods of medical image categorize into: 1. Tamper detection: According to the information analysis of medical image to know whether the image has been tampered. 2. Image recovery: Using the security information which embedded in images to recover the area that had been tampered. This project will analyze the content features of digital X-Ray, computed tomography, magnetic resonance images, nuclear magnetic resonance, ultrasound images those kinds of medical images. During the characteristics of ROI (Region of Interest) and HVS (Human Visual System), we are going to develop high capacity steganography technology for DIOCM medical images. Focusing on the features of DICOM standard, we will improve the spatial domain and frequency domain information hiding methods which only suit for general image formats. Base on the theories of difference expansion, histogram shifting and predict function, we will introduce image tiling and multilayer embedded concepts, different predict algorithms for fragile digital watermarking and reversible steganography technology to make the state-of-the-art method of medical image protection and authentication techniques. Provide some solutions for specific information security problems of digital medical images.

Project ID:PB10501-2699
External Project ID:MOST103-2221-E182-032-MY3