數位影像處理減低 3D 顯示器的重影

Abstract

近年來，由於二維顯示器方面技術的成熟，及一般大眾對於影像娛樂需求的不斷提升，立體顯示器已成為目前各大廠商和研究的主流，期望能藉此提供更真實的視覺感觀效果。為了使得人眼能夠感知到三維立體效果，目前有許多方法被相繼提出，包含了需要戴眼鏡的顯示器，以及不需要戴眼鏡的三維顯示器。然而，不論是哪一種方法，這些顯示器都有著無可避免的漏光，將會使得人眼在觀賞這些影像時，造成極度的不舒服，甚至使得觀賞者無法感知到立體效果。而各家廠商也都致力於解決這個問題，然而，在硬體上的結構總有它的極限，漏光仍然是無可避免的。 因此本篇論文反其道而行，我們利用這種必然會漏光的機制，透過影像處理去改變輸出的影像，使得顯示出的圖像與漏光補償，讓觀賞者能夠看到與原本相同的影像。其中包括不同的影像演算法，並搭配了模擬平台去模擬不同狀況下的影像處理結果。最後透過相機拍下實際經過影像處裡前後的圖像，比較之後，可以發現使用影像處理後，影像品質可以大幅提升，突破硬體上的極限。而最重要的是，此類的影像處理方法，可以運用在幾乎所有的三維顯示器上，並使得影像品質獲得良好的改善。

In recent years, 3D displays can provide more realistic and exciting sensation and extra depth information which is lost in a 2D display. Therefore, it attracts more and more industrial attention. In the past decade, many methods were proposed to implement auto-stereoscopic and stereoscopic displays. But almost all of them were suffered from crosstalk issue. The ghost image which was caused by crosstalk will let the observers feel uncomfortable, and even make they disable to fuse the two image together. So far, there are many methods were proposed to reduce crosstalk. However, they all have their own drawbacks, such as lower luminance, the need of extra device and only for one specific technology. Therefore, we proposed different kind of Digital Crosstalk Reduction methods to suppress optical crosstalk in 3D displays without using any extra hardware components. Furthermore, these methods can be applied to almost all of the 3D display technologies. In this thesis, different kinds of image processing method will be describe in detail, and the results will be captured by a camera to show the difference before and after process.