Reversible Data Hiding Algorithm For Image With High Payload And High Fidelity

In order to effectivly prevent the digital information from being tampered illegally,information hiding becomes an essential part of the research in information security.As the branch of information hiding,reversible data hiding is widely applied in areas including medical,military and legal images.This is because these images not only require to hide information for ownership protection,but also require that the original image is able to be recovered without any distortion at the extraction stage.If any slight distortion is introduced to the recovered images,then it would lead to mistakes in medical treatment or legal evidence.To this end,the topic reversible data hiding,has both the theoretical basis and the practical values.In this paper,we propose two reversible data hiding methods by adopting new prediction mechanism and smoothness measurement.(1)We propose a reversible data hiding algorithm based on k-pass pixel value ordering.In this paper,a new method is proposed to calculate the local complexity of a block.For a block,the pixels with property of embedding invariance within it,coupled with its neighborhood are employed to compute its local complexity.In this way,the accuracy in measuring the local complexity is enhanced.Besides,three largest pixels(three smallest pixels)are selected as an embedding unit.When the difference between the largest and third largest pixels is bigger than 1,then we calculate the prediction error by considering the location relationship of the largest and second largest pixels.It helps to improve the prediction performance for pixels and it enables the difference valued 3 to carry 1 data bit.As it turns out,the number of embeddable pixeles is increased dramatically while the amount of shifted pixels is correspondingly decreased.Experimental results demonstrate that the reversibility and superiority of the proposed method.The proposed scheme provides preferable embedding performance.It contributes to higher embedding capacity and lower image distortion.(2)We put forward a reversible data hiding algorithm based on integer transform.Firstly,three types of irregular block partition methods are proposed in this paper.We design different prediction methods for different-sized blocks in order to improve the predictionperformance.Specifically,the center pixel is utilized to predict its neighboring pixels in the prediction process so that the prediction-error list is created.As it turns out,the accuracy in pixel prediction is enhanced and a sharper prediction-error histogram is created due to the high correlation between the center pixel and its neighbors.Secondly,during the process of integer transform,the mean value for each block keeps unaltered before and after the embedding process.To this end,we include the mean value into the calculation of the local complexity in order to improve the accuracy in local complexity measurement.Finally,two-layer embedding is adopted to optimize the embedding performance.The experimental results indicate that the proposed algorithm is effective.It provides superior embedding performance.