Shot reconstruction degree: a novel criterion for key frame selection

被引：91

作者：

Liu, TY

Zhang, XD ^{[1
]}

Feng, H

Lo, KT

机构：

[1] Tsinghua Univ, Dept Elect Engn, Multimedia Signal Proc Lab, Beijing 100084, Peoples R China

[2] City Univ Hong Kong, Dept Comp Engn & Informat Technol, Hong Kong, Hong Kong, Peoples R China

[3] Hong Kong Polytech Univ, Dept Elect & Informat Engn, Ctr Multimedia Signal Proc, Hong Kong, Hong Kong, Peoples R China

来源：

PATTERN RECOGNITION LETTERS | 2004年 / 25卷 / 12期

关键词：

video content analysis; video summarization; key frame selection;

D O I：

10.1016/j.patrec.2004.05.020

中图分类号：

TP18 [人工智能理论];

学科分类号：

081104 [模式识别与智能系统]; 0812 [计算机科学与技术]; 0835 [软件工程]; 1405 [智能科学与技术];

摘要：

In this paper, the authors treat the key frame extraction problem from a viewpoint of shot reconstruction. As a result, a novel criterion called shot reconstruction degree (SRD) is proposed for key frame selection based on the degree of retaining motion dynamics of a video shot. Compared with the widely used fidelity criterion, the key frame set produced by SRD can better capture the detailed dynamics of the shot. Using the new SRD criterion, a novel inflexion-based key frame selection algorithm is developed. Simulation results show that the new algorithm results in good performance in terms of both fidelity and shot reconstruction degree. (C) 2004 Elsevier B.V. All rights reserved.

引用

页码：1451 / 1457

页数：7

共 12 条

[1]

Content-based representation and retrieval of visual media: A state-of-the-art review [J].