您好, 访客   登录/注册

基于多尺度分解的色调映射算法

来源:用户上传      作者: 胡庆新 陈云 方静

  摘要:针对高动态范围(HDR)图像显示于普通显示设备的问题,提出一种新的基于多尺度分解的色调映射(TM)算法。首先利用局部边缘保留(LEP)滤波器对HDR图像进行多尺度分解,有效平滑了图像的细节同时保留了突出的边缘;根据分解后各层的特点和压缩的要求,提出一个带参数的动态范围压缩函数,通过变化参数以便压缩图像的粗尺度层并增强细尺度层,从而压缩图像的动态范围并增强细节;最后重组各层并恢复颜色,所得到的映射后图像具有良好的视觉效果。实验结果证明,该方法在自然度、结构保真度和整体的质量评价上都要优于Gu等(GU B, LI W J, ZHU M Y, et al. Local edgepreserving multiscale decomposition for high dynamic range image tone mapping [J]. IEEE Transactions on Image Processing, 2013, 22(1): 70-79)和Yeganeh等(YEGANEH H, WANG Z. Objective quality assessment of tonemapped images [J]. IEEE Transactions on Image Processing, 2013, 22(2): 657-667)提出的方法,同时也避免了局部色调映射算法所普遍存在的光晕效应。该算法可以用于HDR图像的色调映射。
  关键词:色调映射;局部边缘保留滤波器;高动态范围图像;多尺度分解;自然度
  中图分类号: TP391.413
  文献标志码:A
  Abstract: A new Tone Mapping (TM) algorithm based on multiscale decomposition was proposed to solve a High Dynamic Range (HDR) image displayed on an ordinary display device. The algorithm decomposed a HDR image into multiple scales using a Local EdgePreserving (LEP) filter to smooth the details of the image effectively, while still retaining the salient edges. Then a dynamic range compression function with parameters was proposed according to the characteristics of the decomposed layers and the request of compression. By changing the parameters, the coarse scale layer was compressed and the fine scale layer was boosted, which resulted in compressing the dynamic range of the image and boosting the details. Finally, by restructuring the image and restoring the color, the image after mapping had a good visual quality. The experimental results demonstrate that the proposed method is better than the algorithm proposed by Gu et al. (GU B, LI W J, ZHU M Y, et al. Local edgepreserving multiscale decomposition for high dynamic range image tone mapping [J]. IEEE Transactions on Image Processing, 2013, 22(1): 70-79) and Yeganeh et al. (YEGANEH H, WANG Z. Objective quality assessment of tonemapped images [J]. IEEE Transactions on Image Processing, 2013, 22(2): 657-667) in naturalness, structural fidelity and quality assessment; moreover, it avoids the halo artifacts which is a common problem existing in the local tone mapping algorithms. The algorithm can be used for the tone mapping of the HDR image.
  Key words: Tone Mapping (TM); Local EdgePreserving (LEP) filter; High Dynamic Range (HDR) image; multiscale decomposition; naturalness
  0引言
  高动态范围(High Dynamic Range, HDR)图像是一类设计用于存储真实世界的亮度值的图像,它通常使用浮点数来表示每个颜色分量。而要想在显示器和打印机上显示HDR图像,则需要一些映射以压缩HDR图像,称为色调映射(Tone Mapping, TM)或者色调重建(tone reproduction)。TM算法最初利用Stevens定律[1]来完成动态范围压缩,之后涌现出了一大批TM算法,通常可将它们分为3类。   第1类是基于感知的TM算法,空间上可分为全局算法和局部算法。全局TM算法[2-3]是对整幅图像的每个像素应用相同的TM曲线,映射时每个像素独立于其相邻像素。根据人类视觉系统(Human Visual System, HVS)的特性,大多数全局算法都利用了非线性映射函数[4-5]。全局算法的优点是结构简单且运算速度快,但是会导致结果图像颜色、细节、对比度和纹理等信息的丢失。因此,人们提出了局部算法以便弥补全局算法的不足,同时在对比度的压缩方面和局部细节的保留方面表现突出。文献[6]中的方法改善了细节但是引入了光晕等视觉上的瑕疵;文献[7]中提出了基于概率模型的TM算法,利用一种概率模型对动态范围压缩过程进行建模,并将其转换为一个能量最小化问题,保留更多的细节并避免了光晕,但有时会产生严重的偏色。
  第2类是基于梯度的TM算法,该类算法依赖衰减大强度的梯度同时保留小幅的震动的思想。文献[8]根据各向异性扩散的思想,利用基于偏微分方程的分层方法,以压缩大强度的梯度;文献[9-10]使用边缘保留的双边滤波器来保留细节以便获得更好的结果;文献[11-12]在梯度域上先对亮度图像进行多尺度衰减,再对新梯度图像通过求解偏微分方程来获得低动态范围图像,这些方法也经常容易产生类似光晕等视觉上的瑕疵;文献[13-14]提出了基于梯度的多尺度分解滤波器,在保留细节和边缘的同时进行图像的压缩。
  第3类是交互式TM算法[15-16],用于为用户选择的特定亮度域实现满意的结果。
  本文利用局部边缘保留(Local EdgePreserving, LEP)滤波器[14]对输入HDR图像进行多尺度分解,提出一个带参数的动态范围压缩函数,通过变化参数对各层进行不同程度的压缩以便获得较好的自然度[17]和局部对比度,所得映射后图像具有良好的视觉效果,避免了偏色和光晕等视觉瑕疵。
  4结语
  本文利用文献[14]中的LEP滤波器对HDR图像进行多尺度分解,针对该算法压缩函数易产生失真和光晕现象的缺陷,提出了一种更好的带参数的压缩函数。该压缩函数的两个参数分别控制了TM后图像的平均亮度和对比度,从而可以根据用户的需求和实际图像的特点进行调整,以便得到最好的输出图像。从结果中可以看出,本文算法无论在局部对比度和自然度上都要优于其他算法,且避免了光晕等现象。
  参考文献:
  [1]STEVENS S. On the psychophysical law [J]. Psychological Review, 1957, 64(3): 153-181.
  [2]TUMBLIN J, RUSHMEIER H. Tone reproduction for realistic images [J]. IEEE Computer Graphics and Applications, 1993, 13(6): 42-48.
  [3]CHIU K, HERF M, SHIRLEY P, et al. Spatially nonuniform scaling functions for high contrast images [C]// GI93: Proceedings of Graphics Interface 1993. Toronto: [s.n.], 1993: 245-253.
  [4]TUMBLIN J, TURK G. Low Curvature Image Simplifiers (LCIS): a boundary hierarchy for detailpreserving contrast reduction [C]// SIGGRAPH 2000: Proceedings of the 27th Annual Conference on Computer Graphics. New York: ACM Press, 2000: 83-90.
  [5]PATTANAIK N S, TUMBLIN J, YEE H, et al. Timedependent visual adaptation for realistic image display [C]// SIGGRAPH 2000: Proceedings of the 27th Annual Conference on Computer Graphics. New York: ACM Press, 2000: 47-54.
  [6]REINHARD E, STARK M, SHIRLEY P, et al. Photographic tone reproduction for digital images [J]. ACM Transactions on Graphics, 2002, 21(3): 267-276.
  [7]SONG M, WANG H, CHEN C, et al. Tone mapping for high dynamic range image using a probabilistic model [J]. Journal of Software, 2009, 20(3): 734- 743.(宋明黎,王惠琼,陈纯,等.基于概率模型的高动态范围图像色调映射[J].软件学报,2009,20(3):734- 743.)
  [8]TUMBLIN J, TURK G. LCIS: a boundary hierarchy for detailpreserving contrast reduction [C]// SIGGRAPH99: Proceedings of the 26th Annual Conference on Computer Graphics. New York: ACM Press, 1999: 83-90.   [9]DURAND F, DORSEY J. Fast bilateral filtering for the display of highdynamicrange images [C]// SIGGRAPH 2002: Proceedings of the 29th Annual Conference on Computer Graphics and Interactive Techniques. New York: ACM Press, 2002: 257-266.
  [10]KUANG J, JOHNSON G, FAIRCHILD M. iCAM06: a refined image appearance model for HDR image rendering [J]. Journal of Visual Communication and Image Representation, 2007, 18(1): 406-414.
  [11]FATTAL R, LISCHINSKI D, WERMAN M. Gradient domain high dynamic range compression [J]. ACM Transactions on Graphics, 2002, 21(3): 249-256.
  [12]MANTIUK R, MYSZKOWSKI K, SEIDEL H. A perceptual framework for contrast processing of high dynamic range images [J]. ACM Transactions on Applied Perception, 2006, 3(3): 286-308.
  [13]FARBMAN Z, FATTAL R, LISCHINSKI D, et al. Edgepreserving decompositions for multiscale tone and detail manipulation [J]. ACM Transactions on Graphics, 2008, 27(3): 1-10.
  [14]GU B, LI W J, ZHU M Y, et al. Local edgepreserving multiscale decomposition for high dynamic range image tone mapping [J]. IEEE Transactions on Image Processing, 2013, 22(1): 70-79.
  [15]PARDO A, SAPIRO G. Visualization of high dynamic range images [J]. IEEE Transactions on Image Processing, 2003, 12(6): 639-647.
  [16]LISCHINSKI D, FARBMAN Z, UYTTENDAELE M, et al. Interactive local adjustment of tonal values [J]. ACM Transactions on Graphics, 2006, 25(3): 646-653.
  [17]YEGANEH H, WANG Z. Objective quality assessment of tonemapped images [J]. IEEE Transactions on Image Processing, 2013, 22(2): 657-667.
转载注明来源:https://www.xzbu.com/8/view-5972370.htm