High-Dynamic-Range Images Compression Algorithm Based On TMO Variation Method

Compared with traditional images,high dynamic range image(HDRI)can have a broader dynamic range,which can fully display the real scene and bring people an excellent viewing experience.However,in order to save such a high dynamic range,the storage space occupied by a single pixel of high dynamic range image is large,which brings great challenges to the transmission and storage of HDR content.This paper aims to propose a HDR image compression algorithm to solve the problem of high memory occupation and high transmission cost of HDR image.There are two major contributions of this paper.First,contemporary high-dynamicrange(HDR)images contain hidden information that can be extracted and exploited to create extra scenes.This phenomenon is carefully examined and analyzed,and a simple TMO is designed to amplify and extract the hidden information so that it is not lost from compression.Secondly,for HDR image compression,a TMO variation method is proposed,which has the following properties.(1)The proposed HDR image compression algorithm based on the TMO variation method can compromise and thus control the image quality between the HDR and the LDR versions.A notable application of this compromising feature is that the hidden information in the original HDR image can be purposely preserved.(2)The proposed HDR image compression algorithm is practical,with no complex optimization employed in the encoding process.(3)Most importantly,the compression efficiency is significantly improved over the 2-layer compression algorithm JPEGXT.Our test results show that the average objective rate savings over JPEGXT are-65% for the tone mapped HDR images and-80% for the linearly recorded HDR images,which are respectively equivalent to an extra 3-and 5-times compression.