利用固定圖案比較的電視節目廣告移除方法

Abstract

影像壓縮技術的運用在生活中隨處可見，如在有限的頻寬中，用影像壓縮技巧來逹到減少資料位元數傳送的目地，或將數小時的電視節目經過壓縮儲存在有限的記憶裝置中。對於喜愛看電視的人而言，如能將節目中的廣告移除，不但能減少儲存空間，且能減少觀看的時間。 基於以上的原因，本論文想要在影像壓縮的範疇下，以固定圖案比較法移除電視節目廣告的部份，利用兩種不同的偵測方法，有效的偵測出節目中廣告的部份，進而將其移除，以提供使用者觀看及儲存上的便利性。

Applications of video compression technology are now everywhere in our daily life. For example, video compression skills are used to transmit less bits of data within a limited bandwidth and to save several hours of TV programs into a limited memory. For those who like watch TV programs, if we can remove commercials from programs, we can save not only the space of storage but also the time to watch it. Therefore, this thesis tries to design a TV commercial removal system for compressed domain, based on fixed-pattern matching method. This system can efficiently detect and remove commercials from programs by two kinds of detection schemes. It offers users a better way to save or to watch programs.