動設備的使用率，如智慧型手機和平板電腦已在近幾年迅速成長。液晶面板追求更高的解析度、亮度、色彩真實度、和更低的耗電量。傳統RGB彩色濾光片LCD發光效率較低。因此可加入白色子像素以提高亮度和降低能耗。但有許多RGBW子像素渲染（SPR）方法可供選擇，因此需要分別評估，並加以優化。
本研究的目的在於評估不同的SPR參數對影像品質的影響，並根據評估的結果，優化新的SPR演算法。本研究首先利用「等級判定法」評估了灰色成分萃取、彩度增強、子像素佈局、和空間濾波等因子的影響。視覺實驗的結果顯示子像素佈局是提高視覺解析度的關鍵。亮度相關的子像素（白色和綠色）必須均勻分佈在屏幕上。在顏色的指定方面，子像素定址比全像素定址方式更好。如果一個低解析度的輸入影像可以使用超解析度（或反迴旋積）技術提高輸入影像的解析度，應能提昇RGBW-SPR演算法的表現。空間低通濾波以減少色彩混疊失真現象是必要的。彩度需要增豔，以補償加入白色像素導致的彩度下降。在灰色成分萃取方面，取RGB最小值和相對亮度能夠分別維持各主色的彩度和亮度。本研究根據上述實驗結果推導一個新的SPR算法，其空間濾波器透過S-CIELAB進行優化。該演算法經「配對比較」視覺實驗驗證，明顯優於PenTile RGBW和MS-RGBW兩個標竿演算法。對輸入影像作輕微的銳化可進一步提高視覺影像品質。

The use of mobile devices such as smart phone and tablet computer has rapidly increased in recent years. The demand of LCD panels with higher resolution, luminance, color fidelity, and lower power consumption has been constantly growing. Conventional RGB-filter-based LCD suffers from low luminance. White sub-pixels therefore are added to enhance the luminance and so as to reduce power consumption. However there are many ways to perform RGBW Sub-pixel Rendering (SPR) for a LCD.
The aim of this study is first to evaluate the effects of different SPR parameters, such as white extraction, chroma enhancement, sub-pixel layout and spatial filtering, on the perceptual image quality of representative RGBW SPR algorithms using a categorical judgement method. The results show sub-pixel layout is the key to achieve higher image quality. The luminance related sub-pixels (white and green) must be evenly distributed on the screen. Sub-pixel addressing is better than whole-pixel addressing. It will be great if a low resolution input image can be up-sampled to higher resolution using a supper resolution (or de-convolution) technology. Low pass spatial filtering is needed to reduce color aliasing. Chroma need to be enhanced mainly for bright colors. In terms of white extraction, min(R, G, B) and relative luminance can preserve colorfulness and brightness of saturate colors respectively.
The results are used to derive a new SPR algorithm and the spatial filter was optimized by S-CIELAB metric. The proposed algorithm outperforms the benchmark algorithms (PenTile RGBW and MS-RGBW) in the second psycho-visual experiment using paired comparison method. The results also suggest slight image sharpening would further improve its perceptual image quality.

[1] C. H. Brown Elliott, T. L. Credelle, S. Han, M. H. Im, M. F. Higgins, P. Higgins, “Development of the PenTile MatrixTM color AMLCD sub-pixel architecture and rendering algorithms”, SID Journal, Vol. 11(1 ), pp. 89-98 (2003)
[2] L. Wang, Y. Tu, L. Chen, K. Teunissen, and I. Heynderickx, “Trade-off between Luminance and Color in RGBW Displays for Mobile-phone Usage”, SID Digest, Vol. 38(1), pp. 1142-1145 (2007)
[3] C. H. Brown Eillott, M. F. Higgins, S. Hwang, S. Han, A. Botzas, B.-S. Hsu, M. Im, and S. Nishimura, “PenTile RGBW Color Processing”, SID Digest, Vol.39(1), pp. 1112-1115 (2008)
[4] C.-C. Lai and C.-C. Tsai, “A Modified Stripe-RGBW TFT-LCD with Image-Processing Engine for Mobile Phone Displays”, IEEE, Consumer Electronics, Vol. 53, pp. 1628-1633 (2007)
[5] K.-Y. Hung, C.-C. Pei, C.-J. Hu, T.-C. Yang, “Manipulation image processing algorithmic technology to realize 1.8” RGBW transflective TFT-LCDs with adjustable colour gamut”, Displays, Vol. 29, pp. 526-535 (2008)
[6] C. H. Brown Elliott and M. F. Higgins, “Sub-pixel Rendering Filters for High Brightness Sub-pixel Layouts”, US patent #7248268 (2007)
[7] M. A. Klompenhouwer, “Flat Panel Display Signal Processing-Analysis and Algorithms for Improved Static and Dynamic Resolution” PhD dissertation, Technische Universiteit Eindhoven, pp. 44-119 (2006)
[8] J. J. Yang, “Color Display Screen and Display Method Thereof”, WO patent #2012171500 (2012)