Clustering multimedia search
View Patent ↗A method for clustering a set of web search results is disclosed. A first signature is compared based at least in part on an analysis of multimedia content associated with a first web search result with a second signature based at least in part on an analysis of multimedia content associated with a second web search result. The first web search result is clustered with the second web search result based at least in part on the comparison of the first signature with the second signature.
1. A method for clustering a set of web search results, comprising:
determining a first video signature based at least in part on an analysis of video content associated with a first web search result;
determining a first audio signature based at least in part on an analysis of audio content associated with the first web search result;
determining a second video signature based at least in part on an analysis of video content associated with a second web search result;
determining a second audio signature based at least in part on an analysis of audio content associated with the second web search result;
comparing the first video signature, the first audio signature, the second video signature, and the second audio signature;
using a processor to cluster the first web search result with the second web search result based at least in part on the comparison; and
labeling the two web search results in the event that:
1) the first and second video signatures are similar and the first and second audio signatures are different or
2) the first and second audio signatures are similar and the first and second video signatures are different.
2. A method as recited in claim 1 , further comprising generating a signature.
3. A method as recited in claim 1 , further comprising reducing the set of web search results based at least in part on an analysis of textual metadata.
4. A method as recited in claim 1 , wherein video content is any non-textual data or metadata.
5. A method as recited in claim 1 , wherein video content includes images, video and audio.
6. A method as recited in claim 1 , wherein each signature has a distance metric between a second signature.
7. A method as recited in claim 6 , wherein the distance metric includes one or a weighted combination of: a Cartesian distance; a Manhattan distance; a Euclidean distance; and a byte difference.
8. A method as recited in claim 6 , wherein clustering includes consolidating web search results with the distance metric below an identical-threshold.
9. A method as recited in claim 6 , wherein clustering includes highlighting web search results with the distance metric below a similar-threshold but above an identical-threshold.
10. A method as recited in claim 1 , wherein each video signature includes a hash value based at least in part on image features.
11. A method as recited in claim 1 , wherein each audio signature includes a hash value based at least in part on audio features.
12. A method as recited in claim 1 , wherein each video signature includes a hash value based at least in part on video features.
13. A method as recited in claim 1 , wherein each video signature includes a recognized face.
14. A method as recited in claim 1 , wherein each video signature includes a recognized logo.
15. A method as recited in claim 1 , wherein each video signature includes a recognized facial feature.
16. A method as recited in claim 1 , wherein clustering labeling includes labeling two web search results with similar video signatures and different audio signatures as commentary.
17. A method as recited in claim 1 , wherein labeling includes labeling two web search results with similar audio signatures and different video signatures as remixes.
18. A method as recited in claim 1 , further comprising precomputing signatures before forming the set of web search results.
19. A method as recited in claim 18 , further comprising finding video that sounds like a web search result.
20. A method as recited in claim 18 , further comprising finding video that looks like a web search result.
21. A method as recited in claim 18 , further comprising finding audio that sounds like a web search result.
22. A system for clustering a set of web search results, comprising:
a data store configured to store signatures of web search results; and
a processor coupled to the data store and configured to:
determine a first video signature based at least in part on an analysis of video content associated with a first web search result;
determine a first audio signature based at least in part on an analysis of audio content associated with the first web search result;
determine a second video signature based at least in part on an analysis of video content associated with a second web search result;
determine a second audio signature based at least in part on an analysis of audio content associated with the second web search result;
compare the first video signature, the first audio signature, the second video signature, and the second audio signature;
cluster the first web search result with the second web search result based at least in part on the comparison; and
label the two web search results in the event that:
1) the first and second video signatures are similar and the first and second audio signatures are different; or
2) the first and second audio signatures are similar and the first and second video signatures are different.
23. A computer program product for clustering a set of web search results, the computer program product being embodied in a non-transitory computer readable medium and comprising computer instructions for:
determining a first video signature based at least in part on an analysis of video content associated with a first web search result;
determining a first audio signature based at least in part on an analysis of audio content associated with the first web search result;
determining a second video signature based at least in part on an analysis of video content associated with a second web search result;
determining a second audio signature based at least in part on an analysis of audio content associated with the second web search result;
comparing the first video signature, the first audio signature, the second video signature, and the second audio signature;
clustering the first web search result with the second web search result based at least in part on the comparison; and
labeling the two web search results in the event that:
1) the first and second video signatures are similar and the first and second audio signatures are different; or
2) the first and second audio signatures are similar and the first and second video signatures are different.