一种基于多尺度DoG滤波器的高动态范围的映射方法

Multi-scale DoG Filter Based High Dynamic Range Image Rendering Method

提出了利用多尺度DoG滤波器模拟人眼视觉的边缘相阻机制,计算出人眼观察各像素的适应性亮度,并对图像各适应性亮度下的色度进行适应性调整;通过计算光感知细胞在各适应性亮度关系下的相对视觉感知信息,提出可保持视觉空间中各维颜色的相对视觉感知信息的映射算法,从而实现了基于多尺度DoG滤波器的高动态范围图像映射。该算法可以在保留高动态范围图像中场景的纹理细节和颜色的前提下,使高动态范围图像能真实地显示在动态范围较低的显示设备或纸张上。实验测试说明,该算法能够正确地再现高动态范围图像原有的场景信息。

A high dynamic range（HDR） image rendering method based on multi-scale DoG filter is proposed.This method compresses dynamic range of HDR images for display devices with limited dynamic range.Meanwhile,this method can preserve fine detail and most color of their original HDR images.In this paper,a multi-scale DoG filter is used to mimic lateral inhibition of HVS to compute the adapted luminance levels for each pixel.Then the adapted luminance levels are combined with a chromatic adaptation model to adjust colors caused by change of illuminant.Finally,rendition of HDR image can be obtained by preserving the relative visual information of each pixel which is computed by the photoreceptor response curves based on the corresponding adapted luminance level.The experimental results demonstrate that the proposed method has good performance,such as rendition of visible fine details of original HDR images.