近來許多研究人員對於如何增進處理家庭影片的效率越來越感興趣。影片內容分析的一個重要技術就是主要物體偵測。偵測影片中的主要物體就是尋找影片中主要物體的位置和所佔據的區域，相當於決定影片的感興趣區域(Region of Interest, ROI)。在後續應用之前導入主要物體偵測的優點有兩項：一、降低處理全畫面的計算量。隨著畫面解析度的提高，從以往的VGA (640*480) 到現今的HD (1920*1080)，畫面處理的運算量以平方的倍率增加。若直接以全畫面的大小來做處理，則運算量將會相當可觀，相對而言，所需的時間會提高許多。事先決定ROI的區域能縮小後續應用所需處理的畫面大小，有效降低運算量並節省處理時間；二、避免萃取到沒有價值的特徵，例如背景的特徵。在此我們提出了一個決定家庭影片中感興趣區域（Region of Interest，ROI）的新方法；四種特徵被萃取並用來計算各別的主要特徵圖像，這四種特徵分別是Multi-Scale Contrast (MSC)、Center-Surround Color Difference (CSCD)、Color Spatial Distribution (CSD) 和相鄰影片框架之間的運動。實驗結果顯示所提出的方法與目前其他方法有一樣的效能。

Recently, researchers are getting more interest of how to improve the efficiency of processing home videos. One of the key techniques of content analysis is salient objects detection. Detecting the salient objects in videos is to find the location and occupied region of the salient object, or the region of interest (ROI) in videos. To reduce the computation loading for processing the whole frame, and avoiding extracting useless features, such as features from the background, we propose a new method to determine the region of interest (ROI) in videos; four salient features are extracted and computed for each salient feature map, Multi-Scale Contrast (MSC), Center-Surround Color Difference (CSCD), Color Spatial Distribution (CSD) and motion. Experimental results show that the proposed method has similar performance compared with prior works.

[1] L. Itti, C. Koch, and E. Niebur. “A Model of Saliency- Based Visual Attention for Rapid Scene Analysis”. IEEE Trans. on PAMI, 20(11):1254–1259, 1998.
[2] Y.F. Ma and H.J. Zhang, “Contrast-Based Image Attention Analysis by Using Fuzzy Growing.” Proceedings of the Eleventh ACM International Conference on Multimedia, 2003, pp. 374–381.
[3]R. Achanta, F. Estrada, P. Wils, and S. susstrunk. “Salient Region Detection and Segmentation”. In ICVS’08.
[4] T. Liu, J. Sun, N. Zheng, X.Xiao, and H.Shum. “Learning to Detect a Salient Objects”. In CVPR, pages 1–8, 2007.
[5] S. Frintrop, M. Klodt, and E. Rome. “A Real-Time Visual Attention System Using Integral Images”. In International Conference on Computer Vision Systems (ICVS'07), March 2007.
[6] Frintrop, S. “VOCUS: A Visual Attention System for Object Detection and Goal-directed Search”. PhD thesis Rheinische Friedrich-Wilhelms-Universit¨at Bonn Germany (2005). Published 2006 in Lecture Notes in Artificial Intelligence (LNAI), Vol. 3899, Springer Verlag Berlin/Heidelberg.
[7]Crow, F. C. “Summed-Area Tables for Texture Mapping”. Proceedings of SIGGRAPH 18 (3, 1084) 207–212.
[8] F. Porkili. “Integral Histogram: A Fast Way to Extract Histograms in Cartesian Spaces”. In CVPR, pages 829– 836, 2005.
[9] I. Pirnog, C. Opera and C. Paleologu. “Image Content Extraction Using a Bottom-Up Visual Attention Model”. In ICDS, 2009
[10] T. Liu, N. Zheng, W. Ding, and Z. Yuan. “Video Attention: Learning to Detect A Salient Object Sequence”. In ICPR, pages 1-4, 2008.
[11] M. Tian, S. Wan, Y. Ji and L. Yue. “Salient Objects Detection in Time Sequenced Images”. In IJCNN, 2008.
[12] C. C. Ho, W. H. Cheng, T. J. Pan and J. L. Wu. “A User-Attention Based Focus Detection Framework and Its Applications”. In ICICS-PCM, 2003.
[13] M. Guironnet, N. Guyader, D. Pellerin and P. Ladret. “Spatio-Temporal Attention Model for Video Content Analysis”. In IEEE, 2005.
[14] Y. Hu, X. Xie, W. Y. Ma, L. T. Chia and D. Rajan. “Salient Region Detection using Weighted Feature Maps based on the Human Visual Attention Model”. In PCM, 2004.
[15] K. Rapantzikos, N. Tsapatsoulis, Y. Avrithis and S. Kollias. “Bottom-Up Spatiotemporal Visual Attention Model for Video Analysis”. In IET Image Processing, 2007.

[16] P. Zhaoyu, L. Pinging and L. Changjiu. “Automatic Detection of Salient Object based on Multi-Features”. In IEEE, 2008.
[17] Bur A., Wurtz P., Muri R.M. and Hugli H.. “Dynamic Visual Attention: Competitive Versus Motion Priority Scheme”. In ICVS, 2007.
[18] B. D. Lucas and T. Kanade. “An Iterative Iimage Registration Technique with an Application to Stereo Vision”. In Proceedings of DARPA Image Understanding, 1981, pp. 121–130