The Secure Model Based On Reversible Data Hiding And Encryption For Medical Images

With the development of scientific technology and the explosive growth of data,big data,the Internet and cloud computing have emerged at the historic moment.At the same time,it has rapidly promoted the development of data transmission,data mining,and data storage.However,data security has been facing a challenge.Nowadays,big data exists in various forms,such as text,images and so on.What is more,images are commonly used carriers for data storage and transmission and they carry information on people's physical conditions.As medical technology develops,people are facing data security issues while enjoying the convenience brought by telemedicine and medical sharing.Hence,this thesis proposed a secure model based on reversible data hiding and encryption for private protection and image security.1.Propose a secure model based on adaptive texture-classification reversible data hiding and plaintext encryption.Reversible data hiding based on adaptive texture classification is proposed,due to the key of medical image is focused on the high texture region.Privacy data is embedded into corresponding medical image for visual quality improvement and privacy protection.For images' security,the idea of plaintext encryption is proposed to transform medical images into the approximate image of target image according to the target image,which reduces the attention to the encrypted medical image and the possibility of being attacked.Experiments have shown the proposed reversible data hiding method improves the visual quality of images obviously.The performance of contrast enhancement by the proposed method is superior to other reversible data hiding methods.Plaintext encryption can protect the medical image content and reduce the possibility of being attacked effectively.2.Propose a security model based on contrast-enhancement reversible data hiding and homomorphic encryption.The lesion of medical images is an important basis for the clinical diagnosis.Aiming atthe lesion area,reversible data hiding based on lesion extraction is proposed.The patient's privacy data is preferentially embedded in the lesion area and the rest data is embedded in the non-lesion area to improve visual quality and the embedding rate.For resisting attacks on exhaustive analysis and statistical analysis,homomorphic encryption based on chaotic mapping is proposed to encrypt medical images to improve image security.Experiments have shown the proposed reversible data hiding method enhances the contrast of the lesion region significantly.The performance of contrast enhancement by the proposed method is better than other reversible data hiding methods.Homomorphic encryption not only can protect images' content,but also can resist the brute-force attack and statistical analysis attack.