採用區塊斜率樣本之消失點偵測技術的二維至三維影像轉換系統

A 2D-to-3D Image Conversion System Using Block Slope Pattern Based Vanishing Point Detection Technology

陳柏瑋(Po-Wei Chen)

指導教授 ： 范育成

國立臺北科技大學/電資學院/電腦與通訊研究所/碩士(2011年)

摘要

近年來三維立體電視(Three Dimensional Television; 3DTV)儼然已經成為新一代多媒體應用上的新技術，市場上也陸續有3DTV的產品開始販售，而目前尚有許多影像無法直接在3DTV顯示立體影像，主要的原因是舊有的影像資訊是由單一平面攝影機或相機來拍攝，這些影像屬於二維(Two Dimension; 2D)的平面影像，二維影像中深度資訊是沒有被記錄的，因此，近年來許多研究單位開始進行二維平面影像轉換成三維立體影像的研究，其中一個研究方法就是利用單一張二維平面影像進行影像分析，演算法分析完成後產生相對應於二維平面影像的深度圖(Depth Map)，而深度圖乃是3DTV成像的重要元素。本論文中將提出一個完整且具可行性的方法來進行重建二維平面影像之深度圖，我們所提出的方法是以區塊為基礎(Block Based)的方式針對二維平面影像進行分析，在影像分析過程中，利用事先建立完整的斜率樣本(Slope Pattern)來辨識區塊中的線段斜率，再利用線段在四種方位上交會的權重值來判斷消失點位置，而消失點即為該影像中深度最遠的位置，最後利用消失點的位置重建影像深度圖，我們所提出的方法專注於基本影像架構元素上的分析，可避免複雜的數學運算，適用於超大積體電路晶片設計上。

關鍵字

立體電視 ； 深度圖 ； 消失點 ； 斜率樣本

並列摘要

Three dimensional television (3DTV) is clearly a new generation of multimedia applications in recent years. At present, there are many existing images can not be displayed directly on 3DTV. The main reason is that these images are captured by a single camera. These images are two dimensional (2D) images which depth information was not recorded. Therefore, many research institutes focus on develop the 2D to 3D conversion algorithm. One of the research methods operates on single monoscopic image. When the conversion process is complete, the algorithm generates the depth map corresponding to the original 2D image. This paper presents a robust and reliable method for the depth map reconstruction. Our proposed method is based on block based approach to analysis the monoscopic image. In the image analysis process, the prior establishment of slope pattern was used to identify the slope of line segment. The Line segments will intersect on vanishing point, and the vanishing point is the furthest position of the monoscopic image. Finally, the vanishing point will be used to reconstruct depth map. We focus on fundamental image structure analysis instead of complex mathematical operation, and the method is suitable for VLSI design.

並列關鍵字

3DTV ； Depth Map ； Vanishing Point ； Slope Pattern

參考文獻

[3] What is stereoscopic vision [online]. http://www.tech-faq.com/

[4] 邱鐘慶，應用於三維立體電視之深度影像重建系統晶片設計與實現，碩士論文，國立台北科技大學，電腦與通訊研究所，台灣台北，2010。

[5] Liang Zhang and Tam W.J., “Stereoscopic Image Generation Based on Depth Images for 3D TV,” IEEE Trans. on Broadcasting, vol. 51, no. 2, pp. 191-199, June 2005.

[6] Akar G. B., Tekalp A. M., Fehn C., and Civanlar M.R., “Transport Methods in 3DTV—A Survey,” IEEE Trans. on Circuits and Systems for Video Technology, vol. 17, no. 11, pp. 1622-1630, Nov. 2007.

[7] Redert A., de Beeck M. O., Fehn C., Ijsselsteijn W., Pollyfeys M., Van Gool L., Ofek E., Sexton I., and Surman P., “Advanced Three-Dimensional Television System Technologies,” in Proc. the First International Symposium on 3D Data Processing Visualization and Transmission, Padova, Italy, Jun. 19-21, 2002, pp. 313-319.