基於顏色與SURF特徵之球場廣告看板內容辨識與計次系統

運動是跨國界與文化的全民活動，因此越來越多企業透過贊助運動賽事增加曝光度，藉由電視廣告以及比賽場館的廣告看板，來強化消費者對贊助商的印象。球場內部大型的廣告看板是球場內最直接醒目的廣告實體，加上透過電視鏡頭的轉播攝影不定時的拍攝到畫面，廣告看板的內容也將呈現給每位電視前的觀眾。因此球場內所有的廣告看版都有極高的曝光率，廣告效果十分驚人，而從前係以人工方式計算廣告次數及時間，常因數據不精確造成廣告商質疑。若能夠利用電腦視覺的技術，有效統計出每次轉播的廣告曝光時間，不僅能節省許多人力，更有利於球團與廣告商進行價碼協商。本論文以顏色與SURF(Speeded-Up Robust Features)為主要特徵，分別作為在影片中偵測與辨識廣告看板的依據，透過顏色的濾除之後再以SURF特徵做辨識，搭配畫面差異(Frame Difference)的偵測，可以大幅降低運算時間，而辨識部分利用特徵描述子的拉氏信號(Sign of the Laplacian)、特徵向量的距離以及合成向量的主要方向比對特徵點的相似度，根據實驗的結果，我們得到Precision Rate 是98.32%、Recall Rate是98.45%，且平均一秒可以執行27.1張畫面，無論在辨識率或是執行速度，都有不錯的效果。

關鍵字

廣告看板； SURF ；特徵描述；物件辨識

並列摘要

Every year sponsors spend large amounts of money on sports marketing, a large portion of which is spent on placement of billboards, banners, and other advertising media positioned. Therefore counting the number of times of the billboards in a sport video is of great important in the marketing and sponsoring sector. In this paper, we describe a system for recognition and counting the number of times of billboards appearing in sports videos. We propose a compact representation of billboards and video frame content based on color and SURF feature points. This representation can be used to robustly detect and recognize billboards as they appear in a variety of different sports video types. Experimental results are provided, along with an analysis of the precision and recall. Results show that our proposed technique is efficient and effectively detects and recognize billboards.

並列關鍵字

Billboard ； SURF ； Feature Description ； Object Recognition

參考文獻

[1] J. F. Canny, “A Computational Approach to Edge Detection,” IEEE Trans. Pattern Analysis and Machine Intelligence, 1986, pp. 679-698.

[4] K. Mikolajczyk and C. Schmid, “Indexing based on Scale Invariant Interest Points,” International Conference on Computer Vision, 2001, pp. 525-531.

[5] D. G. Lowe, “Object Recognition from Local Scale-Invariant Features,” International Conference on Computer Vision, vol. 60, 2004, pp. 1150-1157.

[6] Y. Ke and R. Sukthankar, “PCA-SIFT: A More Distinctive Representation for Local Image Descriptors,” Proceedings Conference Computer Vision and Pattern Recognition, 2004, pp. 511-517.

[9] J. J. Koenderink, “The Structure of Images, Biological Cybernetics,” vol.50, no. 5, 1984, pp. 363–370.

國際替代計量

基於顏色與SURF特徵之球場廣告看板內容辨識與計次系統

全文下載

主題瀏覽