在科技發達的今日，已經有許多資訊傳播媒介已經不再是單純使用傳統文字傳遞訊息，而是改用更方便的裝置如電視牆及跑馬燈等。當人們面對這麼多資訊時難免有遺漏的狀況出現，而在這些遺漏的資訊中有許多資訊對我們來說可能是相當重要且緊急的。因此如何幫助人們在這資訊過於快速的時代中完整重現遺漏的資訊便成為一項重要的課題。
　　由於跑馬燈能夠在有限的版面上快速呈現大量的文字訊息且造價相對便宜，因此相當普及，但因為跑馬燈的顯示範圍有限，一般情況下無法同一時間顯示完整的文字訊息，而需要不斷更新訊息，導致人們有一定的機會遺漏文字資訊。由於現今智慧型手機的普及，因此本論文提出一個以手持裝置為基礎的方法，用以擷取跑馬燈上的文字資訊來幫助人們避免遺漏掉跑馬燈上的重要訊息。首先在偵測階段我們利用一些前景偵測的方法來找出跑馬燈在影片中的位置，接著在重組階段利用本論文提出的過濾機制，將影片中不同樣的文字過濾出來，並將其重組來得到完整的文字訊息。
　　實驗部份我們用了兩種方法來對偵測階段做探討，分別為偵測跑馬燈位置的精準度以及跑馬燈在影片中被偵測到的準確度；在重組階段的探討中，我們採用了正確擷取文字的字數來計算重組階段的正確率。實驗結果顯示，在影片晃動的情況下無論是在白天或是晚上皆能有良好的正確率。

Nowadays, traditional text-based message in conveying information can no longer meet the demand of information broadcasting. Instead, people convey important information via TV walls or marquees (i.e. scrolling texts) benefiting from the emerging of mature technology. Encountering such vast information conveyed by scrolling texts, people may miss some important information due to the inherent scrolling nature. How to help people catching the information hence becomes an important issue to be pursued.
In this thesis, a novel system is proposed to extract texts displayed on an electronic marquee, especially when the input videos are captured using hand-held devices. The proposed system consists of two stages including detection stage and reconstruction stage. In the detection stage, the position of a marquee in each frame is located by utilizing the Gaussian Mixture Model and optical flow information. In the reconstruction stage, a LDP based text-filtering method is designed to retrieve complete text information.
Experiments were conducted to verify the validity of the proposed system. Among them, two experiments were conducted to demonstrate the accuracy of the detected marquee region. As to another experiment, it was conducted to demonstrate the performance in the construction stage by counting how many words are correctly retrieved. Experimental results show that the proposed system works well in capturing scrolling texts both in day and night.

[1]　M. Grundmann, V. Kwatra, I. Essa. “Auto-Directed Video Stabilization with Robust L1 Optimal Camera Paths,” Computer Vision and Pattern Recognition (CVPR) 2011. pp. 225 – 232.
[2]　徐聖哲,數位影像穩定技術及其應用,交通大學博士論文,2010
[3]　Ｙ. Matsushita, E. Ofek, X. Tang, H.Y. Shum. “Full-frame Video Stabilization”, Computer Vision and Pattern Recognition (CVPR) 2005. pp. 50 – 57.
[4]　C. Liu, C. Wang, R. Dai. ” Text Detection in Images Based on Unsupervised Classification”, International Conference on Document Analysis and Recognition (ICDAR) 2005. pp. 610 – 614.
[5]　 B.M. Saturnino, L.A. Sergio, G.J. Pedro, G.M. Hilario, L.F. Francisco. “Road-Sign Detection and Recognition Based on SUpport Vector Machines”, Intelligent Transportation Systems, 8(2), 2007. pp. 264 – 278.
[6]　C. Yi, Y. Tian, A. Arditi. “Portable Camera-Based Assistive Text and Product”, Mechatronics, 19(3), 2014. pp. 808 – 817.
[7]　Haque, M. Murshed, M. Paul, M. “A Hybird Object Detection Technique from Dynamic Background Using Gaussian Mixture Models”, Multimedia Signal Processing, 2008, pp. 915 – 920.
[8]　 M.R. Lyu, J. Song, M. Cai, “A Comprehensive Method for Multilingual Video Text Detection, Localization, and Extraction”, Circuits and Systems for Video Technology,15(2) 2005. pp. 243 – 255.
[9]　陳厚安, 自然場景跑馬燈偵測與完整文具重構,中央大學碩士論文,2011
[10] A.K. Jain, B. Yu, “Automatic Text Location in Images and Video Frames”. Pattern Recognition,31(12),1998, pp. 2055-2076.
[11] Q. Ye, Q. Huang, W. Gao, D. Zhao, “Fast and robust text detection in images and video frames”. Image and Vision Computing,23(6), 2005, pp. 565-576.
[12] K. Jung, K.I. Kim, and A.K. Jain, “Text information extraction in images and videos: A survey”. Pattern Recognition, 37(5), 2004, pp.977-997.
[13] P. Shivakumara, T.Q. Phan and C.L. Tan, “A Robust Wavelet Transform Based Technique for Video Text Detection”, International Conference on Document Analysis and Recognition (ICDAR), 2009, pp. 1285-1289.
[14] B.Epshtein, E. Ofek, Y. Wexler, “Detecting Text in Natural Scenes with Stroke Width Transform” Computer Vision and Pattern Recognition (CVPR) 2010, pp. 2963-2970.
[15] J. Zhang, R. Kasturi,“Text Detection Using Edge Gradient and Graph Spectrum”, International Conference on Pattern Recognition (ICPR), 2010, pp. 3979-3982.
[16] C. Jung, Q. Liu, J.Kim, “A stroke filter and its application to text localization”, Pattern Recognition Letters,30(2), 2009, pp. 114-122.
[17] C.M. Wang, K.C. Fan, C.T. Wang, “Estimating Optical Flow by Integrating Multi-Frame Information”, Journal of Information Science and Engineering, 24(6), 2008, pp.1719-1731.