The purpose of Computer Vision is to understand the methods by which humans\rprocess visual information and likewise to create computer algorithms similar to these\rprocesses. Through careful observation, a computer algorithm was developed to mimic\rhow humans recognize logos in television commercials. After visual analysis of\rnumerous commercial sequences, it was hypothesized that the key frames (frames in\rwhich the logo resides) could be found using the intersection of color histograms; the\rlogo region could be found using the edge density within the key frames; and the logo\rcould be identified utilizing a correlation method with a database of stored logos, scaled\rto different levels using Bilinear Interpolation.\rColor histograms were implemented using one-dimensional arrays with 24 bins;\rkey frames were determined by calculating the intersection between consecutive frames’\rcolor histograms. The edge density was calculated by convolving the key frame with\rthe number of edge pixels within a 21X21 area. The identification of the logo was\rdetermined by computing the Sum of Square Differences between the logo region and\rthe database of logos on different scales; SSD values were normalized for different\rscales.\rThe algorithm was tested on 14 different sequences and determined the key frame\rwith 80% accuracy. By segmenting the sequence into two key frames, the algorithm\rgenerated 93% accuracy. The algorithm also identified the logo region with 93%\raccuracy. The identification of the logo yielded anomalous results. These data suggest\rthat motion between consecutive frames in commercial segments decreases around the\rdisplay of the logo. They also suggest that the logo region has the most visible edges\rwithin the key frame.\rFuture study includes a complete overhaul of the logo recognition algorithm. The\rcorrelation algorithm (SSD) does not work accurately enough to be used. Therefore, the\rnext step is possibly to look at the edge information about the key frames. As the Canny\ralgorithm determines the edges of an image, it has to determine the direction (or\rorientation) of the edges. Therefore, a proposed study includes utilizing an edge\rorientation histogram of the database of the logos and the key frames. This would mean\rthat the algorithm would identify the logo in the key frames by matching edge\rorientation histograms.
「為配合國家發展委員會「推動ODF-CNS15251為政府為文件標準格式實施計畫」，以及 提供使用者有文書軟體選擇的權利，本館檔案下載部分文件將公布ODF開放文件格式， 免費開源軟體可至LibreOffice 下載安裝使用，或依貴慣用的軟體開啟文件。」