- 學位論文
場色序液晶顯示器色分離抑制之超快速最佳化演算法
Ultra-Fast Optimization Algorithm for Color Breakup Suppression in FSC LCDs
摘要
場色序式液晶顯示器（FSC-LCDs）由於不需要使用彩色濾光片，相較於傳統的液晶顯示器擁有低成本、高透光率的優點，然而此種顯示器存在一重要問題－色分離現象。色分離現象成因乃人眼與影像之間有一相對速度，子畫面在視網膜上無法完全重疊，在邊緣會有顏色錯位的現象而經視覺暫留產生色分離，此現象嚴重影響場色序式液晶顯示器之影像品質，一般來說有三個方向來抑制該現象－增加畫面應答速率、減少顏色錯位部分色差、動態畫面補償。一般的場色序式液晶顯示器顯示速率大部分為180（三個子畫面）赫茲或240赫茲（四個子畫面）。 由於目前高速率的液晶難以發展與動態畫面補償容易造成過高的演算法與硬體成本，本論文採用上述第二種抑制色分離現象的方式（減少顏色錯位部分色差）並且分別在180赫茲與240赫茲的應用上提出了演算方式。針對減少顏色錯位部分色差，240赫茲的應用上採用了將能量集中在主色畫面的方法，使得該主色畫面的影像盡可能與欲顯示之圖片相近，然後再用另外三子畫面補足剩餘的顏色；而180赫茲上的應用則是兼顧了兩個面向以減少顏色錯位部分色差：(1)集中能量在主色畫面、(2)調整背光使得三子畫面擁有相近的背光。 實驗結果顯示，240赫茲的場色序式液晶顯示器色分離抑制之超快速最佳化演算法能夠比以往的方法集中能量，目前可達平均80%以上的能量集中度，並且在個人電腦上的執行速度小於0.1秒；而180赫茲的場色序液晶顯示器色分離抑制之超快速最佳化演算法則是能夠同時集中能量、減少三子畫面背光色差並且控制失真率平均在3%以下。
並列摘要
FSC-LCDs (field-sequential-color LCDs) have the merits of low cost and high luminous efficiency due to color filter-less structure. However, FSC-LCDs are facing a serious artifact, color break-up (CBU), which degrades image quality. CBU is perceived when a relative velocity exists between the object on the screen and the observer’s eyes. Sub-frames are incapable of totally overlapping on retinas, and CBU occurs owing to photogene. There are three ways to suppress CBU, increasing frame rate, reduction of color difference and dynamic compensation. Most studies about CBU suppression in LCDs focus on 180Hz (3 fields) and 240Hz (4 fields) applications. Because of the bottleneck of high-speed Liquid Crystal development and the unaffordable cost of dynamic compensation, this thesis chooses to do reduction of color difference and presents algorithm s for 180Hz and 240Hz application. The proposed 240Hz algorithm concentrates most energy on the first field and makes the first field to resemble the source image as much as possible. Then, the other three fields are residual frames that maintain the image fidelity. On the other hand, the proposed 180Hz algorithm not only concentrates most energy on the first field but also makes the back light of all 3 fields to be similar. Experiment results verify the average energy concentration ratio is over 80% for proposed 240Hz algorithm. And the execution speed is less 0.1 second per image on PC. For 180Hz, the average energy concentration ratio is over 55% and the average distortion is less then3%.