2004
DOI: 10.1007/978-3-540-27814-6_49
|View full text |Cite
|
Sign up to set email alerts
|

Feature Based Cut Detection with Automatic Threshold Selection

Abstract: There has been much work concentrated on creating accurate shot boundary detection algorithms in recent years. However a truly accurate method of cut detection still eludes researchers in general. In this work we present a scheme based on stable feature tracking for inter frame differencing. Furthermore, we present a method to stabilize the differences and automatically detect a global threshold to achieve a high detection rate. We compare our scheme against other cut detection techniques on a variety of data … Show more

Help me understand this report

Search citation statements

Order By: Relevance

Paper Sections

Select...
3
1
1

Citation Types

0
32
0
5

Year Published

2005
2005
2017
2017

Publication Types

Select...
5
1
1

Relationship

0
7

Authors

Journals

citations
Cited by 35 publications
(37 citation statements)
references
References 9 publications
(8 reference statements)
0
32
0
5
Order By: Relevance
“…The results of the comparisons are shown in Table 4. It can be seen that our method produces better results than threshold oriented visual feature-based and edge histogram-based methods [12][13][14], respectively. This results from the fact that the motion effects are better handled by our method.…”
Section: Classification and Experimental Resultsmentioning
confidence: 99%
See 1 more Smart Citation
“…The results of the comparisons are shown in Table 4. It can be seen that our method produces better results than threshold oriented visual feature-based and edge histogram-based methods [12][13][14], respectively. This results from the fact that the motion effects are better handled by our method.…”
Section: Classification and Experimental Resultsmentioning
confidence: 99%
“…A threshold-based automatic cut detection method is introduced in [12]. The method uses visual features for representing cut candidates and, according to these features, threshold values are estimated.…”
Section: Related Workmentioning
confidence: 99%
“…Most systems, including this work, incorporate two main lowlevel processing steps: (a) shot boundary detection [6,18,23], which is to find the boundaries between stitched shots, and (b) camera motion estimation within shots [2,11,14,19,24,25] to further decompose shots into finer sub-shot units based on evolving camera motion types.…”
Section: Related Workmentioning
confidence: 99%
“…Shot boundary detection is believed to be largely solved [18]; we adopt [23]. For background (BG) camera motion estimation, we extend [14,24] to estimate three P/T/Z camera motion parameters from KLT tracks while simultaneously separating the tracks into FG/BG groups.…”
Section: Related Workmentioning
confidence: 99%
See 1 more Smart Citation