Procedings of the British Machine Vision Conference 2004 2004
DOI: 10.5244/c.18.20
|View full text |Cite
|
Sign up to set email alerts
|

Activity Based Video Content Trajectory Representation and Segmentation

Abstract: A novel approach is developed to segment continuous CCTV recordings according to the activities captured in the scene. This approach differs from previous approaches which are mostly based on shot change detection and shot grouping. Video content is represented by constructing a cumulative multi-event histogram over time. An on-line segmentation algorithm is then proposed to detect breakpoints in the video content, which is more robust to noise and computationally much more efficient compared to existing on-li… Show more

Help me understand this report

Search citation statements

Order By: Relevance

Paper Sections

Select...
3
1
1

Citation Types

0
24
0

Year Published

2006
2006
2011
2011

Publication Types

Select...
4
2

Relationship

1
5

Authors

Journals

citations
Cited by 19 publications
(24 citation statements)
references
References 16 publications
0
24
0
Order By: Relevance
“…In a not-too-busy scenario, there are often non-activity gaps between two consecutive behaviour patterns which can be utilised for activity segmentation. In the case where obvious non-activity gaps are not available, an on-line segmentation algorithm proposed in [22] can be adopted.…”
Section: Behaviour Pattern Representationmentioning
confidence: 99%
“…In a not-too-busy scenario, there are often non-activity gaps between two consecutive behaviour patterns which can be utilised for activity segmentation. In the case where obvious non-activity gaps are not available, an on-line segmentation algorithm proposed in [22] can be adopted.…”
Section: Behaviour Pattern Representationmentioning
confidence: 99%
“…In a not-too-busy scenario, there are often non-activity gaps between two consecutive behaviour patterns which can be utilised for activity segmentation. In the case where obvious non-activity gaps are not available, an on-line segmentation algorithm proposed in [26] can be adopted. More specifically, surveillance video contents are firstly represented holistically in space and over time based on discrete events detected automatically in the scene resulting in a high-dimensional video content trajectory (more details about the discrete scene events follows).…”
Section: Behaviour Representationmentioning
confidence: 99%
“…More specifically, surveillance video contents are firstly represented holistically in space and over time based on discrete events detected automatically in the scene resulting in a high-dimensional video content trajectory (more details about the discrete scene events follows). The break points on the trajectory correspond to video content changes and can be detected using the on-line algorithm proposed in [26].…”
Section: Behaviour Representationmentioning
confidence: 99%
“…The lengths of these video segments were within the range of 127 to 3210 frames. In our experiment, a scene event based method proposed in [20] was adopted for feature extraction, which resulted in each image frames being represented as a 8 dimensional feature vector. The problem to be solved here is to automatically determine K, the number of hidden states which corresponds to the number of video segments.…”
Section: Surveillance Video Segmentationmentioning
confidence: 99%