Security Model for Medical Image Data Hidden Using a New Digital Watermark and Steganography Technique , LL-INS-TU1A, 13019222, Tokuo Umeda, "/> <span style="font-size: 12px;">Security Model for Medical Image Data Hidden Using a New Digital Watermark and Steganography Technique </span>
RSNA 2013 

Abstract Archives of the RSNA, 2013


LL-INS-TU1A

Security Model for Medical Image Data Hidden Using a New Digital Watermark and Steganography Technique 

Scientific Informal (Poster) Presentations

Presented on December 3, 2013
Presented as part of LL-INS-TUA: Informatics - Tuesday Posters and Exhibits (12:15pm - 12:45pm)

Participants

Tokuo Umeda PhD, Presenter: Nothing to Disclose
Akiko Okawa MD, RN, Abstract Co-Author: Nothing to Disclose
Tsutomu Gomi PhD, Abstract Co-Author: Nothing to Disclose
Kenta Miwa RT, Abstract Co-Author: Nothing to Disclose
Shuji Yamamoto PhD, Abstract Co-Author: Nothing to Disclose
Taku Yashima RT, Abstract Co-Author: Nothing to Disclose

PURPOSE

The number of images included in a single image series is increasing with the development of multi-detector computed tomography (MDCT). In addition, the application of information and communication technology (ICT) techniques, such as tele-image-reading and external archiving, has been introduced in the medical field. Therefore, there is increasing risk at the time of archiving images and transmission of electronic patient records (EPR). We present information hiding system developed using digital watermarking and steganography technologies.

METHOD AND MATERIALS

1. Digital watermarking technology Chest CT images of 512×512×16 bits (67 slices) were used. The EPR data, the hash value of the regions of interest (ROI), name of the institute, and the data of the patient support system were hidden in the regions of non-interest (RONIs) in a chest CT image series in the Digital and Communication in Medicine (DICOM) format. 2. Steganography technology A body CT image series of 512×512×16 bits (100 slices) was used for steganography. These CT images were stored in a subfolder after 7-Zip compression. This folder was then embedded in the cover image of a scene photograph. The cover image with the embedded images was then transmitted to other medical institutions.

RESULTS

When part of the ROIs was altered during transmission, the hash value decoded from the received cover image had a different value from that before cover image transmission. The structural similarity (SSIM) and the peak signal to noise ratio (PSNR) of the watermarked image with 4000 words embedded were 0.99 and 65.3 dB, respectively. In addition, when the medical information was embedded in the low-bit plane, such as the first- and second-bit plane, the radiologist was unable to identify the embedded information. In our technology, there was no change in the image capacity of CT images or cover image before and after embedding.

CONCLUSION

Using digital watermarking and steganography technologies, we developed a medical information hiding system that ensures the copyright of the images, and protects privacy and safety of the EPR. Both technologies are applicable not only to CT images but to all digital images.

CLINICAL RELEVANCE/APPLICATION

The proposed scheme would allow central management of medical images and EPR values, and thus facilitate efficient handling of medical information.

Cite This Abstract

Umeda, T, Okawa, A, Gomi, T, Miwa, K, Yamamoto, S, Yashima, T, Security Model for Medical Image Data Hidden Using a New Digital Watermark and Steganography Technique .  Radiological Society of North America 2013 Scientific Assembly and Annual Meeting, December 1 - December 6, 2013 ,Chicago IL. http://archive.rsna.org/2013/13019222.html