System and method for auto content recognition
View Patent ↗System and method for automatically recognizing media contents comprise steps of capturing media content from the Internet and/or devices, extracting fingerprints from captured contents and transferring to the backend servers for identification, and backend servers processing the fingerprints and replying with identified result.
1. A method of auto content recognition by VDNA (Video DNA, here DNA simply stands for identifier) fingerprints of media contents for video, image and audio contents, said method comprising:
a process of VDNA fingerprint extraction to extract characteristic values of each frame of image or audio from said media contents on front-end devices,
a process for fingerprints data stream to be divided into pieces of fingerprint data with any time interval of timestamps as separate VDNA fingerprint data files, and
a process for identification server to contain location information where said media contents are played, time information when said media contents are played and total amount of people who pay attention to said media contents to analyze popularity of said media contents in different areas at different times.
2. The method as recited in claim 1 , wherein said characteristic values include waves of sound, levels of color and shapes of waves for audio and video of said media content.
3. The method as recited in claim 1 , wherein if said VDNA fingerprints are generated continuously on said front-end devices, playing timestamps will be provided along with said VDNA fingerprints.
4. The method as recited in claim 1 , wherein said media contents going to play at future time can be predicted by said VDNA fingerprints as soon as current playing contents are recognized, and so advertisements embedded inside said media contents can be predicted.
5. The method as recited in claim 4 , wherein said advertisements can be changed by content owners by pushing new advertisements to said front-end devices at the predicted time.
6. The method as recited in claim 1 , wherein said media contents playing on target sources including television broadcasting and Internet sharing will be recognized automatically by said VDNA fingerprints, and content owners only need to assign resources to specify said target sources and monitor said media contents.
7. The method as recited in claim 1 , wherein said VDNA fingerprints are sent to said identification server together with captured timestamps via network.
8. The method as recited in claim 1 , wherein said identification server will wait until fingerprint data is enough to identify, and identification result will be combined with captured timestamps with earlier identification history to achieve more accurate results.
9. The method as recited in claim 1 , wherein said VDNA fingerprints allow to identify said media contents that are not exactly the same as the sample master media contents wherein small changes like watermarks, station logos, borders, banners and capturing noises are allowed and have only little influence in identification process.
10. The method as recited in claim 1 , wherein said front-end device can store said VDNA fingerprints data with timestamps in its storage in offline mode without network connection to said identification server.
11. A system of auto content recognition by VDNA (Video DNA, here DNA simply stands for identifier) fingerprints of media contents for video, image and audio contents, said system comprising:
a module for VDNA fingerprint extraction to extract characteristic values of each frame of image or audio from said media contents on front-end sub-systems,
a module for fingerprints data stream to be divided into pieces of fingerprint data with any time interval of timestamps as separate VDNA fingerprint data files, and
a module of identification server to contain location information where said media contents are played, time information when said media contents are played and total amount of people who pay attention to said media contents to analyze popularity of said media contents in different areas at different times.
12. The system as recited in claim 11 , wherein said characteristic values include waves of sound, levels of color and shapes of waves for audio and video of said media content.
13. The system as recited in claim 11 , wherein if said VDNA fingerprints are generated continuously on said front-end sub-systems, playing timestamps will be provided along with said VDNA fingerprints.
14. The system as recited in claim 11 , wherein said media contents going to play at future time can be predicted by said VDNA fingerprints as soon as current playing contents are recognized, and so advertisements embedded inside said media contents can be predicted.
15. The system as recited in claim 14 , wherein said advertisements can be changed by content owners by pushing new advertisements to said front-end sub-systems at the predicted time.
16. The system as recited in claim 11 , wherein said media contents playing on target sources including television broadcasting and Internet sharing will be recognized automatically by said VDNA fingerprints, and content owners only need to assign resources to specify said target sources and monitor said media contents.
17. A method of database generation for auto content recognition by VDNA (Video DNA, here DNA simply stands for identifier) fingerprints of media contents for video, image and audio contents, said method further comprising:
a process of VDNA fingerprint extraction to extract characteristic values of each frame of image or audio from said media contents on front-end devices,
a process for fingerprints data stream to be divided into pieces of fingerprint data with any time interval of timestamps as separate VDNA fingerprint data files,
a process for identification server to contain location information where said media contents are played, time information when said media contents are played and total amount of people who pay attention to said media contents to analyze popularity of said media contents in different areas at different times, and
a process for said identification server to filter subsets of said VDNA fingerprint data from whole fingerprint database using hash and keywords which consumes less resources to compare with a subset of said VDNA fingerprint data than to compare with every entry of said VDNA fingerprint data in said fingerprint database during recognition process.
18. The method as recited in claim 17 , wherein said characteristic values include waves of sound, levels of color and shapes of waves for audio and video of said media content, and metadata of said media contents is collected in back-end database.
19. The method as recited in claim 18 , wherein said metadata is stored in said back-end database of said identification server combined with said VDNA fingerprint data.
20. The method as recited in claim 18 , wherein said metadata can be modified any time by those who have enough privileges to access said back-end database, but said VDNA fingerprints data cannot be changed after extraction.