Video entity recognition in compressed digital video streams
View Patent ↗A method and system for detection of video segments in compressed digital video streams is presented. The compressed digital video stream is examine to determine synchronization points, and the compressed video signal is analyzed following detection of the synchronization points to create video fingerprints that are subsequently compared against a library of stored fingerprints.
1. A method of detecting a video entity in a video stream having a plurality of video frames, the method comprising:
storing a plurality of known fingerprints, wherein each of the known fingerprints includes a calculated feature corresponding to at least one frame of known video content;
receiving the video stream;
defining a plurality of regions of interest in the video frames of the received video stream;
determining a fingerprint of at least one of the video frames based on the defined regions of interest;
comparing the determined fingerprint to the plurality of known fingerprints; and
determining, based on the comparing, that the received video stream contains the video entity.
2. The method of claim 1 , further comprising:
assigning a confidence value based on the comparing, wherein the confidence value represents a threshold level of similarity between the determined fingerprint and the plurality of known fingerprints; and
utilizing the confidence value in determining that the received video stream contains at least one frame of the known video content.
3. The method of claim 1 , wherein the plurality of known fingerprints corresponds to the defined regions of interest.
4. The method of claim 1 , wherein the determining the fingerprint is accomplished based on a comparison of features between at least two of the defined regions of interest.
5. The method of claim 1 , further comprising:
determining the plurality of known fingerprints from the at least one frame of known video content prior to storing the plurality of known fingerprints.
6. The method of claim 5 , wherein the plurality of known fingerprints is determined using the defined plurality of regions of interest.
7. The method of claim 1 , wherein the known fingerprints and the determined fingerprints are statistical parameterized representations of the corresponding video frame.
8. The method of claim 1 , wherein the determining the fingerprint includes calculating a feature of the defined regions of interest of the at least one video frame.
9. The method of claim 8 , wherein the feature used to determine the fingerprint is the same feature as the calculated feature of the known fingerprints.
10. The method of claim 1 , wherein the known fingerprints and the determined fingerprint are based on color coherence vectors of the defined regions of interest.
11. The method of claim 1 , wherein the known fingerprints and the determined fingerprint are based on color histograms of the defined regions of interest.
12. The method of claim 1 , wherein the known fingerprints and the determined fingerprint are based on the relative color of the defined regions of interest.
13. The method of claim 1 , wherein the known fingerprints and the determined fingerprint are based on the brightness of the defined regions of interest.
14. A video entity detection system for detecting a video entity in a video stream having a plurality of video frames, said system comprising:
a memory for storing a plurality of known fingerprints wherein each of the known fingerprints includes a calculated feature corresponding to at least one frame of known video content;
a receiving module configured to receive the video stream;
a fingerprinting module configured to define a plurality of regions of interests in the video frames of the received video stream and to determine a fingerprint of at least one of the video frames based on the defined regions of interest; and
a comparison module configured to compare the determined fingerprints to the plurality of known fingerprints and to determine that the received video stream contains the video entity based on the comparing.
15. The system of claim 14 , wherein the comparison module is further configured to:
assign a confidence value based on the comparing, wherein the confidence value represents a threshold level of similarity between the determined fingerprint and the plurality of known fingerprints; and
utilize the confidence value in determining that the received video stream contains at least one frame of the known video content.
16. The system of claim 14 , wherein the plurality of known fingerprints corresponds to the defined regions of interest.
17. The system of claim 14 , wherein the comparison module determines the fingerprints by comparing the features of at least two of the defined regions of interest.
18. The system of claim 14 , wherein the fingerprinting module is further configured to determine the plurality of known fingerprints from the at least one frame of known video content based on the defined plurality of regions of interest prior to storing the known fingerprints in the memory.
19. The system of claim 14 , wherein the known fingerprints and the determined fingerprints are statistical parameterized representations of the corresponding video frame and the statistical parameterized representations are based on one of color coherence vectors, color histograms, the relative color and the brightness of the defined regions of interest.
20. The system of claim 14 , wherein the fingerprinting module determines the fingerprint of a video frame by calculating a feature of the defined regions of interest of the video frame and the feature is the same feature as the calculated feature of the known fingerprints.