頁籤選單縮合
題名 | Accurate Shot Transition Detection Algorithms |
---|---|
作者姓名(中文) | 林欣政; 李維光; 賴尚宏; | 書刊名 | 影像與識別 |
卷期 | 11:2 民94.06 |
頁次 | 頁65-85 |
專輯 | Semantic Content Analysis for Broadcast TV Program |
分類號 | 312.1 |
關鍵詞 | |
語文 | 英文(English) |
英文摘要 | Most video processing takes require shot transition detection to segment a video into different shots as the first step. In this paper, we propose two different algorithms for shot transition detection. The first on exploits the information extracted from motion and illumination estimation to detect shot boundaries. The second approach is based on a leaning framework that combines different kinds of discriminative video features and uses the SVM classifiers to achieve accurate shot transition detection. We test the proposed algorithms on the TRECVID benchmarking video datasets in our experiment to show that the accuracies of the proposed algorithms are comparable to the best results reported in TRECVID. |
本系統之摘要資訊系依該期刊論文摘要之資訊為主。