Dynamic-range of luminance under standard visual environment, for example among luminance for stars of night sky and for sun-light reaches up to 160 [dB]. Although our visual system can give stable perception against such environment, conventional electronic imaging devices such as digital cameras cannot be able to capture these scenes without lack of information because of the restriction of their dynamicrange. Dynamic-ranges of such conventional electronic imaging devices are only 60 [dB] and lack of dynamic-range yields lacks of information as under exposure or over exposure. Limitation of dynamic-range is crucial problem for widedynamic-range image capturing as vehicle video systems, security cameras and so on. Several methods for such problem were proposed in recent studies. For example, the method for using a series of images with different shutter speed is known as ‘multiple exposure’ can achieve the wide-dynamic-range imaging using pseudo expansion of dynamic-range using conventional electronic imaging device. Recently, this method was implemented as a hardware system combined with a CMOS image sensor to achieve 160 [dB] dynamic-range. The problems for lack of dynamic-range are not only for imaging device but also for image reproduction medium such as displays. Luminance corrections are required to compress the luminance range to fit with image reproduction device's one. This process is known as ‘tone mapping’ which corresponds to compression of the dynamic-range for multiple exposure image to that of image reproduction devices. Tone mapping method by effective luminance remapping is need to perform the WYSIWYG (What You See Is What You Get) image reproduction. This study aims to propose the tone mapping method to achieve the natural contrast reproduction characteristics and to apply it to video sequences. Proposed tone mapping hypothesized that the local luminance adaption to average luminance around each pixel plays an important role to reproduce the spatially-localized contrast with keeping the global luminance variation by simulating the local adaptation mechanisms for the human visual system. In proposed tone mapping, smoothing by a Bilateral filter was applied to suppress the pseudo edges by halo to calculate the luminance of around pixels for each pixel. Additionally, multiresolution representation for spatial frequency was introduced to save the calculation cost. Also, a series of subjective evaluation experiment were performed to investigate the dependency of each parameter for reproduced image quality and the dynamic-range dependencies of the proposed tone mapping. Result showed that proposed tone mapping gave highest quality composed image compare to conventional tone mapping or conventional wide-dynamic-range imaging device.
Toshihiro Toyota, Hiroki Yokomakura, Koichi Kaneko, Akira Usui, Shigeki Nakauchi, "Perceptual Reproduction of Wide-dynamic-range Scene based on Local Adaptation of the Human Visual System" in Proc. IS&T CGIV 2010/MCS'10 5th European Conf. on Colour in Graphics, Imaging, and Vision 12th Int'l Symp. on Multispectral Colour Science, 2010, pp 372 - 379, https://doi.org/10.2352/CGIV.2010.5.1.art00059