#### DMCA

## A shortest path representation for video summarisation (2003)

Venue: | In Proceedings of the 12th International Conference on Image Analysis and Processing |

Citations: | 8 - 1 self |

### Citations

10455 | Introduction to Algorithms
- Cormen, Leiserson, et al.
- 1990
(Show Context)
Citation Context ...the sum of the weights of its constituent edges: ª Ô � �� �� � Ú�s�Ú� � (7) If one or more paths exist from Ú to Ú� the shortest path is defined as the path Ô with the minimum total weight, Ñ�Ò�ª Ô � =-=[3]-=-. The connectivity of a graph can be represented as an adjacency matrix Å in which each element �� � represents the edge between vertices Ú� and Ú�. If there is an edge Ú��Ú� then Å�� � � Ú��Ú� otherw... |

3847 |
Random Sample Consensus: A Paradigm for Model Fitting with Applications to Image Analysis and Automated Cartography
- Fischler, Bolles
- 1981
(Show Context)
Citation Context ...f outliers. Outliers may be present in the data where the motion equation is invalidated or where points correspond to secondary motions. 3 Model Parameter Estimation Robust estimators such as RANSAC =-=[5]-=- are commonly used when least squares estimators can not handle significant numbers of outliers. Using the RANSAC algorithm we can select two points in frame � Òsat random and estimate the model param... |

358 | MLESAC: a new robust estimator with application to estimating image geometry
- Torr, Zisserman
- 2000
(Show Context)
Citation Context ... � Ò � × �Ò� Ü�Ò� Ý�Ò corresponds to the scale, translation in x and translation in y respectively, between the frames � Òsand � Ò . The model parameters are estimated using the robust estimator MSAC =-=[10]-=- which provides good estimates in the presence of outliers. Outliers may be present in the data where the motion equation is invalidated or where points correspond to secondary motions. 3 Model Parame... |

252 | Comparison of video shot boundary detection techniques
- Boreczky, Rowe
- 1996
(Show Context)
Citation Context ...tes both video and audio information and the abstraction is played rather than viewed statically. Whereas many of the approaches to video summarisation rely on the explicit detection of shot changes =-=[1, 7]-=- other techniques have been proposed that commonly transform the image into a lower dimensional space, e.g. using PCA [6]. Typically, frames are then grouped using a clustering algorithm [6] or curve ... |

157 | Comparison of automatic shot boundary detection algorithm
- Lienhart
- 1999
(Show Context)
Citation Context ...tes both video and audio information and the abstraction is played rather than viewed statically. Whereas many of the approaches to video summarisation rely on the explicit detection of shot changes =-=[1, 7]-=- other techniques have been proposed that commonly transform the image into a lower dimensional space, e.g. using PCA [6]. Typically, frames are then grouped using a clustering algorithm [6] or curve ... |

107 | Video summarization by curve simplification
- DeMenthon, Kobla, et al.
- 1998
(Show Context)
Citation Context ...es have been proposed that commonly transform the image into a lower dimensional space, e.g. using PCA [6]. Typically, frames are then grouped using a clustering algorithm [6] or curve simplification =-=[4]-=- and the frame closest to each cluster centroid is chosen as a key frame. Fundamentally, in order to create an efficient index, an algorithm for video abstraction should extract a set of key frames re... |

93 | Video abstracting
- Lienhart, Pfeiffer, et al.
- 1997
(Show Context)
Citation Context ...s problem is to use a video abstract for indexing and retrieval. A video abstract is defined as a short sequence of images, extracted from a longer video whilst still preserving the essential message =-=[8]-=-. The difficulty in composing such an abstract is determining which frames best represent the video contents. A common approach is to segment temporally a sequence into shots, where a shot is defined ... |

40 | Adjustable Filmstrips and Skims as Abstractions for a Digital Video Library
- Christel, Hauptmann, et al.
(Show Context)
Citation Context ...as a sequence of frames captured from a single camera operation, and then select a single representative key frame for each shot. The resulting ordered set of key frames is referred to as a filmstrip =-=[2]-=-. To allow efficient indexing, an abstract must represent the entire video content with as little redundancy as possible. Each key frame should represent a video segment which exhibits consistency in ... |

12 | Visual abstraction of wildlife footage using Gaussian mixture models and the minimum description length criterion
- Gibson, Campbell, et al.
- 2002
(Show Context)
Citation Context ...aches to video summarisation rely on the explicit detection of shot changes [1, 7] other techniques have been proposed that commonly transform the image into a lower dimensional space, e.g. using PCA =-=[6]-=-. Typically, frames are then grouped using a clustering algorithm [6] or curve simplification [4] and the frame closest to each cluster centroid is chosen as a key frame. Fundamentally, in order to cr... |

8 | Thomas: Detection and Classification of Shot Transitions
- Porter, Mirmehdi, et al.
- 2001
(Show Context)
Citation Context ...each shot. In this paper, we propose an algorithm which assumes that a video sequence has already been temporally segmented into individual shots using the edit effect detection algorithm outlined in =-=[9]-=-. Then, estimates of the dominant motion between each frame pair are used to decide when there has been sufficient camera motion to require another key frame. A weighted directed graph is formed for e... |