US6072542A
Expired / Lapsed
US
⏱️ 0.0 years remaining
Automatic video segmentation using hidden markov model
Assignee: Fujifilm Business Innovation Corp
•
Filing: 1997-11-25
•
Granted: 2000-06-06
•
Expires: 2017-11-25
Abstract
Abstract Detection of video shot boundaries using a Video Segmenting Hidden Markov Model to model the sequence of states of a video. The Video Segmenting Hidden Markov Model determines the state sequence based on feature values. Using Hidden Markov Model techniques allows for automatic learning and use of multiple features including motion vectors, audio differences and histogram differences, without the need for manual adjustments of these thresholds.
Inventors
- 👤 Lynn D. Wilcox
- 👤 John S. Boreczky
Bibliographic Details
Application Number
—
Jurisdiction
US
Data Source
google_patents
Last Synchronized
2026-11-09 18:58
CPC Classifications
G; G06; G06V; G06V20/00; G06V20/40; G11; G11B; G11B27/00; G11B27/10; G11B27/19; G11B27/28; H; H04; H04N; H04N5/00; H04N5/14; H04N5/147
Claims Analysis
Total Claims: 21
Claim 1
Independent Claim
71 words
A method for segmenting a video, comprising: extracting at least one feature from the video; comparing the at least one extracted feature to a model having a plurality of states; inferring a sequence of the states of the video, based on the comparison of the at least one extracted feature and the states and probabilities of moving between the states; and segmenting the video based on the inferred sequence of states.
Claim 2
Independent Claim
71 words
A method for segmenting a video, comprising: extracting at least one feature from the video; comparing the at least one extracted feature to a model having a plurality of states; inferring a sequence of the states of the video, based on the comparison of the at least one extracted feature and the states and probabilities of moving between the states; and segmenting the video based on the inferred sequence of states.
Claim 3
Dependent Claim
47 words
The method of claim 1, wherein a plurality of features are extracted from the video, the plurality of extracted features being compared to a corresponding plurality of model parameters of the states of the model, the extracted features and the model parameters being independent of each other.
Claim 4
Dependent Claim
33 words
The method of claim 2, wherein the sequence of states is inferred based on finding a best-fit between the extracted features and the model parameters and the probabilities of moving between the states.
Claim 5
Dependent Claim
16 words
The method of claim 1, wherein the states include shots, transition states and camera movement states.
Claim 6
Dependent Claim
18 words
The method of claim 4, wherein the probabilities of moving between the states are equal, for all states.
Claim 7
Dependent Claim
26 words
The method of claim 4, wherein the transition states comprise cuts, fades and dissolves and wherein the camera motion states comprise shots with pans and zooms.
Claim 8
Dependent Claim
18 words
The method of claim 1, wherein the features comprise gray-level histogram difference levels, motion vectors and audio differences.
Claim 9
Independent Claim
56 words
A method of training a video segmenting model comprising: indexing a training video; manually assigning boundaries to states of the training video, the states comprising shots, camera motion and transition states; extracting at least one feature from the training video; determining feature values for the at least one feature; and computing distributions of the feature values.
Claim 10
Independent Claim
56 words
A method of training a video segmenting model comprising: indexing a training video; manually assigning boundaries to states of the training video, the states comprising shots, camera motion and transition states; extracting at least one feature from the training video; determining feature values for the at least one feature; and computing distributions of the feature values.
Claim 11
Independent Claim
64 words
An apparatus for segmenting a video into states, comprising: an input device that inputs a digitized video; a processor, coupled to the input device, that processes the digitized video to extract a feature value and to produce a distribution of the feature value; and a module that segments the digitized video given a set of feature values and a video segmenting hidden markov model.
Claim 12
Independent Claim
64 words
An apparatus for segmenting a video into states, comprising: an input device that inputs a digitized video; a processor, coupled to the input device, that processes the digitized video to extract a feature value and to produce a distribution of the feature value; and a module that segments the digitized video given a set of feature values and a video segmenting hidden markov model.
Claim 13
Dependent Claim
34 words
The apparatus of claim 9, wherein the processor extracts a plurality of features, and produces a corresponding plurality of distributions, and wherein the processor determines the best fit based on the plurality of features.
Claim 14
Dependent Claim
16 words
The apparatus of claim 10, wherein each of the features are independent from the other features.
Claim 15
Dependent Claim
16 words
The apparatus of claim 9, wherein transition probabilities are equal, for all movement among the states.
Claim 16
Dependent Claim
90 words
A data structure residing in a memory, the data structure comprising: a model having a plurality of states; a first probability associated with each state for each of the at least one feature; second probabilities associated with moving among the plurality of states; and an algorithm for determining an optimum sequence of states, based on observed feature values for the at least one feature and the first probabilities associated with the states, and the second probabilities wherein the model is used to segment a video into a sequence of states.
Claim 17
Dependent Claim
90 words
A data structure residing in a memory, the data structure comprising: a model having a plurality of states; a first probability associated with each state for each of the at least one feature; second probabilities associated with moving among the plurality of states; and an algorithm for determining an optimum sequence of states, based on observed feature values for the at least one feature and the first probabilities associated with the states, and the second probabilities wherein the model is used to segment a video into a sequence of states.
Claim 18
Dependent Claim
17 words
The data structure of claim 13, wherein the states comprise cuts, fades, dissolves, shots, pans and zooms.
Claim 19
Dependent Claim
28 words
The data structure of claim 14, wherein the algorithm determines the optimum sequence of states based on the probabilities associated with the states for a plurality of features.
Claim 20
Dependent Claim
19 words
The data structure of claim 15, wherein the features comprise gray-level histogram difference levels, motion vectors and audio differences.
Claim 21
Dependent Claim
17 words
The data structure of claim 13, wherein the probabilities are equal for all movement along the states.
References Cited (Backward Citations)
US5649060A
Cited
US5655058A
Cited
US5732146A
Cited
US5786855A
Cited
US5801765A
Cited
US5828809A
Cited
USPTO Annuity & Maintenance Windows
| Annuity Stage | Window Opens | Payment Deadline | Surcharge Period Until | Estimated Fee | Status |
|---|---|---|---|---|---|
| 3.5 Years (Year 4) | 2003-06-06 | 2003-12-06 | 2004-06-06 | $2000.00 | Window Closed / Lapsed |
| 7.5 Years (Year 8) | 2007-06-06 | 2007-12-06 | 2008-06-06 | $3760.00 | Window Closed / Lapsed |
| 11.5 Years (Year 12) | 2011-06-06 | 2011-12-06 | 2012-06-06 | $7700.00 | Window Closed / Lapsed |