A new local dimming algorithm based on the simplex method

详细信息    查看全文

• 作者：Martin Riplinger ; Michael Krause…
• 关键词：LED edge ; lit ; Local dimming ; Optimization ; Real ; time imaging ; High power saving ; Real ; time imaging
• 刊名：Computational Optimization and Applications
• 出版年：2016
• 出版时间：May 2016
• 年：2016
• 卷：64
• 期：1
• 页码：243-263
• 全文大小：936 KB
• 参考文献：1.Albrecht, M., Karrenbauer, A., Xu, C.: Sorted sector covering combined with image condensation-an efficient method for local dimming of direct-lit and edge-lit lcds. IEICE Trans. Electron. 93(11), 1556–1563 (2010)CrossRef
  2.Elble, J.: Computational experience with linear optimization and related problems. PhD thesis, University of Illinois at Urbana-Champaign (2010)
  3.Geiger, C., Kanzow, C.: Theorie und Numerik Restringiererter Optimierungsaugaben. Springer, Berlin (2002)CrossRef
  4.Gill, P., Murray, W., Wright, M.: Numerical Linear Algebra and Optimization. Addison-Wesley Publishing, Boston (1991)MATH
  5.Harris, P.M.: Pivot selection methods of the devex lp code. Math. Progr. 5(1), 1–28 (1973)MathSciNet CrossRef MATH
  6.Jung, T., Albrecht, M., Schäfer, D., Xu, C.: System architecture and FPGA-implementation of the ssc local dimming processor for an edge-lit serial tv. Soc. Inf. Disp. 42, 665–668 (2011)
  7.Karrenbauer, A., Albrecht, M., Xu, C.: Method, system and apparatus for power saving backlight. US Patent 8411116 (2013)
  8.Krause, M., Riplinger, M., Louis, A.K., Xu, C.: A new class of very effective algorithms for local dimming. Submitted (2014)
  9.Natterer, F.: The Mathematics of Computerized Tomography. SIAM, Philadelphia (2001)CrossRef MATH
  10.Pan, P.Q.: Nested pricing for the simplex algorithm: an empirical evaluation. Optimization (2007)
  11.Pan, P.Q.: Efficient nested pricing in the simplex algorithm. Oper. Res. Lett. 36(3), 309–313 (2008)MathSciNet CrossRef MATH
  12.Schäfer, D., Jung, T., Xu, C.: Enhanced local pixel compensation with clipping suppression and global luminance preservation. SID Symp. Dig. Tech. Papers 43, 473–476 (2012)CrossRef
  
• 作者单位：Martin Riplinger (1)
  Michael Krause (2)
  Alfred K. Louis (1)
  Chihao Xu (2)
  
  1. Institute of Applied Mathematics, Saarland University, 66041, Saarbrücken, Germany
  2. Institute of Microelectronics, Saarland University, 66041, Saarbrücken, Germany
  
• 刊物类别：Mathematics and Statistics
• 刊物主题：Mathematics
  Optimization
  Operations Research and Mathematical Programming
  Operation Research and Decision Theory
  Statistics
  Convex and Discrete Geometry
  
• 出版者：Springer Netherlands
• ISSN：1573-2894

文摘

A recently arising important feature of mobile computing is power efficiency. Economical power management allows to increase the battery lifetime of tablets, mobile phones and so on. As the display of those devices yields the largest part of power consumption in the near future, it is necessary to reduce the power consumption of them. This can be achieved by dimming, especially by local dimming, i.e reducing the backlight of the display while the necessary brightness in all pixels has to be satisfied. Mathematically, this problem is modeled in this paper as a linear program, which will be solved by a tuned up version of the popular simplex algorithm. While there are other methods as the sorted sector covering algorithm or the General Dimming Algorithm for local dimming, they do not allow to integrate a-priori information for the solution. For example, restricting the maximum difference of the power of two neighboring LEDs is not possible due to the calculation, while it is a natural demand from the physical point of view. Furthermore, the simplex method allows to integrate smoothness conditions on the solution depending on the second order derivative. Additionally, a problem adapted pivot rule is proposed and the warmstart-strategy of the simplex is elaborated to speed up the proposed algorithm during watching movies. Furthermore, the new method does not only minimize the power consumption, it also aligns the backlight to the image leading to perfectly dimmed images. Numerical examples comparing the different methods substantiate the many advantages of the new method.