Granted Patent B2

US 12,387,497 · App. 18/536,349 · Granted Aug 12, 2025

Inventor: Sean Lawlor (Dollard-des-Ormeaux, CA)

G06V20/52G06V10/764G06V20/40G06V20/47G11B27/3081G11B27/34G06V20/44G08B17/125

Abstract

Claims (51)

1. A computer-implemented media processing method, comprising:

obtaining an original media stream that includes a plurality of time windows;

identifying a video segment of interest that includes a given time window of the plurality of time windows, wherein the given time window is associated with an anomaly state that meets criteria; and

including the video segment of interest and an indication of the anomaly state associated with the given time window in a composite media stream.

2. The method defined in claim 1 , wherein the criteria define an anomalous change in motion including at least one of an anomalous increase in motion and an anomalous decrease in motion.

3. The method defined in claim 1 , wherein the anomaly state associated with the given time window is based on (i) an estimate of statistical parameters about a normal motion level for the original media stream; (ii) instantaneous measures of a current motion level of the given time window; and (iii) utilizing statistical hypothesis testing to determine if the current motion level is statistically normal or abnormal.

4. The method defined in claim 1 , further comprising:

obtaining motion indicators for a plurality of samples of the original media stream, each of the plurality of time windows spanning a subset of the samples;

obtaining the anomaly state for the given time window of the plurality of time windows by:

obtaining estimated statistical parameters for the given time window based on measured statistical parameters characterizing the motion indicators for the subset of the samples in at least one time window of the original media stream that precedes the given time window; and

determining the anomaly state for the given time window based on the motion indicators obtained for the subsets of samples in the given time window and the estimated statistical parameters.

5. The method defined in claim 1 , wherein the original media stream is a video stream originating from a particular camera, and the composite media stream comprises information that identifies the particular camera.

6. The method defined in claim 1 , wherein the video segment of interest further comprises at least one time window that precedes the given time window or follows the given time window.

7. The method defined in claim 1 , further comprising storing the composite media stream in a memory.

8. The method defined in claim 1 , wherein the criteria comprise a threshold, and the original media stream comprises a video stream, wherein identifying the video segment of interest comprises:

obtaining motion indicators for a plurality of samples of the video stream; and

comparing at least one p-value for the motion indicators for samples in the given time window with the threshold to determine whether the anomaly state of the given time window meets the criteria.

9. The method defined in claim 1 , further comprising:

determining a length of a set of time windows associated with the anomaly state that meets the criteria;

calculating a compression ratio based on a length of the original media stream and the determined length; and

displaying the compression ratio on a user interface.

10. The method defined in claim 1 , further comprising generating a media summary that includes the indication of the anomaly state.

11. The method defined in claim 1 , further comprising:

querying a database for a subset of time windows in a desired time range; and

creating a stream summary of the desired time range.

12. A non-transitory computer-readable medium comprising computer-readable instructions which, when executed by a computing device, configure the computing device to carry out a method that includes:

obtaining an original media stream that includes a plurality of time windows;

identifying a video segment of interest that includes a given time window of the plurality of time windows, wherein the given time window is associated with an anomaly state that meets criteria; and

including the video segment of interest and an indication of the anomaly state associated with the given time window in a composite media stream.

13. The non-transitory computer-readable medium defined in claim 12 , wherein the criteria define an anomalous change in motion including at least one of an anomalous increase in motion and an anomalous decrease in motion.

14. The non-transitory computer-readable medium defined in claim 12 , wherein the anomaly state associated with the given time window is based on (i) an estimate of statistical parameters about a normal motion level for the original media stream; (ii) instantaneous measures of a current motion level of the given time window; and (iii) utilizing statistical hypothesis testing to determine if the current motion level is statistically normal or abnormal.

15. The non-transitory computer-readable medium defined in claim 12 , the method further comprises:

obtaining motion indicators for a plurality of samples of the original media stream, each of the plurality of time windows spanning a subset of the samples;

obtaining the anomaly state for the given time window of the plurality of time windows by

obtaining estimated statistical parameters for the given time window based on measured statistical parameters characterizing the motion indicators for the subset of the samples in at least one time window of the original media stream that precedes the given time window; and

determining the anomaly state for the given time window based on the motion indicators obtained for the subsets of samples in the given time window and the estimated statistical parameters.

16. The non-transitory computer-readable medium defined in claim 12 , wherein the original media stream is a video stream originating from a particular camera, and the composite media stream comprises information that identifies the particular camera.

17. The non-transitory computer-readable medium defined in claim 12 , wherein the video segment of interest further comprises at least one time window that precedes the given time window or follows the given time window.

18. The non-transitory computer-readable medium defined in claim 12 , wherein the criteria comprise a threshold, and the original media stream comprises a video stream, wherein identifying the video segment of interest comprises:

obtaining motion indicators for a plurality of samples of the video stream; and

comparing at least one p-value for the motion indicators for samples in the given time window with the threshold to determine whether the anomaly state of the given time window meets the criteria.

19. The non-transitory computer-readable medium defined in claim 12 , the method further comprising:

determining a length of a set of time windows associated with the anomaly state that meets the criteria;

calculating a compression ratio based on a length of the original media stream and the determined length; and

displaying the compression ratio on a user interface.

20. An apparatus comprising:

a non-transitory memory comprising instructions; and

one or more processors in communication with the memory, wherein the one or more processors execute the instructions to:

obtain an original media stream that includes a plurality of time windows;

identify a video segment of interest that includes a given time window of the plurality of time windows, wherein the given time window is associated with an anomaly state that meets criteria; and

include the video segment of interest and an indication of the anomaly state associated with the given time window in a composite media stream.

Assignments (1)

Continuity (6)

References Cited (71)