超大型積體電路與液晶螢幕製程變異診斷及效能最佳化

VLSI and LCD Process Variation Diagnosis and Performance Optimization

李宗昱(Tsung-Yu Li)

指導教授 ： 陳中平

國立臺灣大學/電機資訊學院/電子工程學研究所/碩士(2009年)

https://doi.org/10.6342/NTU.2009.02377

摘要

本篇論文的第一個部份主要是介紹液晶螢幕瑕疵自動化偵測的機制。液晶螢幕為近年來最廣泛使用的一種顯示器，然而在現代化的生產流程之下卻依然必須以人眼作為MURA瑕疵的主要檢測的手段，帶來了高昂的成本與不精確的檢測結果。為了增加效率與提高MURA瑕疵的檢測率，我們提出了一套以MURA邊緣偵測為核心的區域整合MURA自動化檢測機制。
第二個章節則著重於將離散資料擬和為convex function。在得到擬和的convex fucntion之後，就可以透過凸優化數學規劃來得到全域最佳解。凸優化為數學規劃的子領域技術，透過這項技術可以保證對一個最佳化的問題得到近似最佳解。然而要使用凸優化技術，其目標函式與限制式皆必須為convex function而許多最佳化問題之目標函式與限制式通常為離散的資料而且不為convex function，為了能夠將這些最佳化問題轉換為凸優化問題，我們提出了一套對離散資料擬和為convex function的演算法並可以在多項式時間複雜度下以線性規劃完成擬和。

關鍵字

液晶螢幕 ； 效能最佳化 ； 超大型積體電路 ； 邊緣偵測 ； 影像處理 ； 凸函數

並列摘要

The first part of this thesis introduces an automatic LCD defect inspection scheme. The liquid crystal displays (LCDs) have become one of the most widely used displayers in recent years; however, in contemporarily available production processes manual/human inspection of Mura defects is still inevitable, taking tremendous time, costs and occasional inaccuracy or inconsistency. To enhance efficiency, we proposed an alternative automatic Mura defect inspection scheme for region-Mura detection. The core of our solution is to find the boundaries of Mura regions, as we proposed two edge-detecting algorithms to fulfill the task.
The second part of our work focused on resolving a convex function fitting model for discrete data. Having such a fitting convex function can make global optimized solutions much easier to find, because the convex optimization (or convex programming) is a subfield of mathematical optimization which definitely has a global optimized solution. However, for a convex optimization problem, the objective function must be a convex function, and for many applications, the objective data is usually discrete and doesn’t have a convex function. In order to apply the convex optimization technique to the discrete objective data, we therefore propose a new convex fitting approach with linear programming.

並列關鍵字

LCD ； performance optimization ； VLSI ； edge detection ； image processing ； convex function

參考文獻

[2] DunlopFishburn and A.J. (1985). “TILOS: A posynomial programming approach to transistor sizing”. in International Conference on Computer-Aided Design, pp. 326-328.

[3] Kishore KasamsettyKetkar and Sachin S. SapatnekarMahesh. (Apr. 2002). “A New Class of Convex Functions for Delay Modeling and their Application to the Transistor Sizing Problem”. in IEEE Journal of Solid-State Circuits, vol. 37, pp. 521-525.

[4] S. L. ChenH. Chang, L. C. TsaiJ. (2008). “TFT-LCD Mura Defects Automatic Inspection system using Linear Regression Diagnostic Model”. International Journal of Advanced Manufacturing Technology.

[5] S. S. SapatnekarB. Rao, P. M. Vaidya, and S. M. KangV. (Nov. 1993). “An exact solution to the transistor sizing problem for CMOS circuits using convex optimization”. in IEEE Transactions Computer-Aided Design of Integrated Circuits and Systems, vol. 12, no. 11, pp.1621-1634.

[6] S.L. ChenJhouJ.W. (2007). "Automatic Optical Inspection on Mura defect of TFT-LCD".