視訊內容擷取系統之視訊切割與索引方法之研究

Abstract

視訊擷取系統包含索引及查詢兩個主要子系統。在視訊索引建立過程中，視訊切割是第一步且最為重要之步驟。本文提出兩個視訊切割方法:統計圖模糊聚類分群法及統計圖動量不變分群法。此兩種方法均是結合了場景變化偵測法與分群法，並用來解決兩種方法之缺點。統計圖模糊聚類分群法或統計圖動量不變分群法，包含特徵值粹取、分群、及關鍵視訊框選取等三個步驟。首先，粹取出相鄰兩視訊框統計圖之差異值當做特徵值。接著，利用模糊聚類分群法或動量不變分群法將特徵值分為場景變化、可能發生場景變化，及無場景變化等三群。最後，從場景變化、可能發生場景變化之兩群中找出真正場景變化之視訊框，並將視訊切割成數段不同之場景，再從每個場景選出代表視訊框。實驗結果驗證統計圖模糊聚類分群法及統計圖動量不變分群法，均可適用於各種不同型態之視訊。 在視訊索引子系統中，本論文提出利用時間相似度之模數運算視訊內容擷取之架構。在視訊索引建立過程中，分別將一段視訊中兩物件間之所有時序關係，建立時序三重值，並提出一質數分配法，分配質數給所有時序三重值。經由計算每個視訊中所有質數值，即可算出每個視訊之代表值。在查詢過程中，利用連續模數運算方法，比對視訊資料庫中之視訊代表值，即可找出所要查詢之視訊資料。基於質數分配法及連續模數運算方法，本文建立一個實驗性的視訊擷取系統。經由模擬結果之分析，此系統之查訊時間為線性時間。

A video retrieval system consists of two major subsystems for indexing and query, respectively. In the indexing process, video segmentation is the first and the most important step of building a video retrieval system. In this dissertation, we propose two video segmentation methods: the histogram-based fuzzy c-means (HBFCM) clustering algorithm, and the histogram-based moment-preserving (HBMP) clustering algorithm. Each algorithm is the hybrid of the shot change detection approach and the clustering approach, and is designed to overcome the drawbacks of both approaches. The HBFCM or the HBMP clustering algorithms is composed of three phases: the feature extraction phase, the clustering phase, and the key-frame selection phase. In the first phase, differences between color histogram are extracted as features. In the second phase, the fuzzy c-means (FCM) or the moment-preserving is used to group features into three clusters: the shot change (SC) cluster, the suspected shot change (SSC) cluster, and the no shot change (NSC) cluster. In the last phase, shot change frames are identified from the SC and the SSC, and then used to segment video sequences into shots. Finally, key frames are selected from each shot. Simulation results indicate that both of the HBFCM clustering algorithm the HBMP clustering algorithm are robust and applicable to various types of video sequences. In the indexing subsystem, we propose a framework for retrieving video sequences using modular operations on temporal similarity. In the indexing process, the contents (objects) of a video sequence are used to build temporal triples representing temporal relationships between objects. For every temporal triple generated from different video sequences, a prime number is assigned to it by the proposed prime number assignment (PNA) algorithm. For each video sequence, a standing value is generated via its set of prime numbers. In the query process, successive modular operations (SMO) on the standing values are used to retrieve video sequences. Based on the PNA and the SMO, we built an experimental video retrieval system. From the simulation results, we notice that a query can be done in a linear time.