VOL. 10, NO. 14, AUGUST 2015 ISSN 1819-6608 ARPN Journal of Engineering and Applied Sciences © 2006-2015 Asian Research Publishing Network (ARPN). All rights reserved. www.arpnjournals.com 5758 ROBUST IMAGE STEGANOGRAPHY BY EMBEDDING SELECTIVE INTRINSIC MODE FUNCTIONS WITH DISCRETE WAVELET TRANSFORM S. Senthil Kumar 1 and K. Palani Thanaraj 2 1 Department of Electronics and Communication Engineering, Agni College of Technology, Chennai, India 2 Department of Electronics and Instrumentation Engineering, St. Joseph’s College of Engineering, Chennai, India E-Mail: sskpuliyur@gmail.com ABSTRACT Steganography is the method of embedding information in a carrier medium for secure transmission. Steganography enables protection of confidential data that arise in many military and communications systems, online retail and banking systems, medical data transmission etc. This paper focuses on medical image steganography that involves hiding a secret medical image into a cover image thereby preserving patient privacy. In this work improvements to discrete wavelet transform (DWT) based steganography is attempted in different clinical settings. Here we propose to combine DWT with empirical mode decomposition (EMD) at different frequency scales for robust and secure image steganography. Initial step involves decomposing the cover and secret image to predefined approximation level using DWT. Then the approximate secret image is decomposed to intrinsic oscillating modes that contain details at different frequency scales. A selection procedure is initiated in this stage where the user can embed the secret image at different detail level based on the application requirements. The predefined intrinsic modes of secret image are embedded in the cover image by a linear mixing model. Then inverse DWT is applied to reconstruct the stegano image. Performance assessment of the proposed method is carried out (6.5% of payload) with image quality metrics such as peak signal to noise ratio (PSNR), mean Square Error (MSE), maximum absolute error (MAXERR) and ratio of squared norms (L2RAT) are tabulated and compared with DWT based steganography. Our study shows that proposed method can be a robust tool in secure transmission of secret images. Keywords: image steganography, discrete wavelet transform, empirical mode decomposition, intrinsic mode functions. INTRODUCTION Steganography is the science of transmitting confidential data by embedding in a carrier image. The idea behind this is to mainly counter the blind attacks that are on the rise due to recent developments in communication media. Since most of the data transmitted through internet are digital, there are high chances for it to be illegally used or copied. To avoid such illegal incidents, it is necessary to have a secure communication through the internet. Steganography provides secrecy by hiding the secret data into another medium before transmitting it to the end user (Cheddad et al. 2010). This method has been proved to provide a secure communication through the internet. There are various types of Steganography that is currently used for carrying a wide variety of payloads such as text data, image data, voice and video data. Based on the nature of carrier medium steganography can be classified as text steganography, image steganography, video steganography and audio steganography (Bachrach and Shih, 2011). The type of steganography depends on the type of secret data need to be transmitted to the end user. Most of the existing steganography methods rely on two factors: a) the secrecy of the key and b) the robustness of the steganography algorithm (Li et al. 2011) . Robustness is defined as the ability of the hidden data to withstand against intentional and unintentional attacks in noisy transmission environment. In this paper the usage of image steganography for medical applications is investigated. Confidential patient data are stored in Picture Archival and Communication Systems (PACS). This medical data is an electronic media that is transmitted to physician for patient analysis. This is accessible by the patient and physicians anywhere (Ibaida and Khalil, 2013). This added facility comes with a bottleneck of privacy issues. So secrecy of patient data has to be maintained. In this perspective we have taken the studies on hiding a patient anatomical Magnetic Resonance Imaging (MRI) into another cover image thereby blocking unwanted attacks that questions the privacy of patients. For a secure communication, the quality of the stego-image must be high so that the attackers would not find any difference between the stego- image and the cover image. This shows the need for robustness of the steganographic method for a secure communication. There are many techniques used in steganography that can be found in current literature. Information hiding schemes in carrier images can be broadly classified as 1) pixel based methods 2) transform based methods. Least significant bit (LSB) substitution is a common pixel based steganography that hides information in the LSB of the carrier data (Chan and Cheng 2004). It provides high payload capacity with