Media content identification on mobile devices
A mobile device responds in real time to media content presented on a media device, such as a television. The mobile device captures temporal fragments of audio-video content on its microphone, camera, or both and generates corresponding audio-video query fingerprints. The query fingerprints are transmitted to a search server located remotely or used with a search function on the mobile device for content search and identification. Audio features are extracted and audio signal global onset detection is used for input audio frame alignment. Additional audio feature signatures are generated from local audio frame onsets, audio frame frequency domain entropy, and maximum change in the spectral coefficients. Video frames are analyzed to find a television screen in the frames, and a detected active television quadrilateral is used to generate video fingerprints to be combined with audio fingerprints for more reliable content identification.
1. A method for capturing video content for processing, the method comprising:
capturing video content presented on a display screen of a primary device by use of a camera contained in a secondary device;
separating the captured video content from boundaries of the display screen by an analysis application running on the secondary device that determines a quadrilateral shape of the captured video content;
based on determining the quadrilateral shape, overlaying an outline of a quadrilateral on an image of the captured video content on a display screen of the secondary device;
cropping the captured video content to boundaries of the quadrilateral shape of the captured video content;
after cropping the video content, warping the cropped video content to an upright rectangle that adjusts the video content for perspective distortion due to a viewing angle of the camera relative to the display screen of the primary device, wherein a first perspective of the quadrilateral shape is different than a second perspective of the upright rectangle; and
generating video fingerprints of the warped video content by a fingerprint application on the secondary device for transmission to a search function.
2. The method of claim 1 , wherein the primary device is a television (TV) and the secondary device is a smart phone.
3. The method of claim 1 further comprising:
running edge detection on video frames selected from the captured video content to detect edges of the display screen;
running corner detection on results of the edge detection to detect the quadrilateral shape; and
determining the detected quadrilateral shape having appropriate dimensions for the display screen of the primary device.
4. The method of claim 1 wherein:
the analysis application determines a quadrilateral shape having appropriate dimensions for the display screen within selected video frames of the captured video content; and
the method further comprises fingerprinting the separated captured video content of the selected video frames to generate the video fingerprints.
5. The method of claim 4 further comprising:
determining a frame of audio content for a corresponding time period of the selected video frames;
fingerprinting the frame of audio content to generate audio fingerprints; and
combining the video fingerprints with the audio fingerprints from which a selected set of query fingerprints are sent to the search function.
6. The method of claim 1 , wherein the search function is an application that has access to a local search database residing on the secondary device.
7. The method of claim 1 wherein:
the quadrilateral shape represents an active display screen area of the primary device; and
the method further comprises refining the quadrilateral shape by tracking motion of pixel formations from frame to frame of the captured video content.
8. The method of claim 1 further comprising:
the quadrilateral shape represents an active display screen area of the primary device; and
the method further comprises checking brightness and color for the detected quadrilateral shape against the captured video content to improve identification of the active display screen area of the primary device.
9. The method of claim 1 further comprising:
a quadrilateral shape representing an active display screen area of the primary device; and
fingerprinting regions of the active display screen area on the secondary device to generate a plurality of video signatures representing the separated captured video content.
10. The method of claim 1 further comprising:
searching a reference database for reference fingerprints that match the generated fingerprints to identify the separated captured video content; and
upon finding a match to the generated fingerprints, storing the reference fingerprints locally on the secondary device.
11. The method of claim 10 further comprising:
searching for the stored reference fingerprints on the secondary device to track the video content presented on the primary device.
12. The method of claim 1 , wherein the search function operates on a remote search server.
13. The method of claim 1 , wherein the search function operates on the secondary device.
14. The method of claim 1 further comprising:
generating an action signal on the secondary device upon detection of a pre-specified video fragment stored in a reference fingerprint database; and
in response to the action signal, performing a pre-specified action based on local search results.
15. The method of claim 14 , wherein the pre-specified action is presenting information regarding the local search results to a user of the secondary device.
16. A processing system for capturing video content from a primary device for processing on a secondary device, the system comprising:
a storage unit on the secondary device having a camera to capture video content; and
a processing unit on the secondary device configured to use the camera:
to capture video content presented on a display screen of the primary device,
to separate the captured video content from boundaries of the display screen utilizing an analysis application running on the secondary device that determines a quadrilateral shape of the captured video content,
to overlay an outline of the quadrilateral shape on an image of the captured video content on a display screen of the secondary device based on determining the quadrilateral shape,
to crop the captured video content to boundaries of the quadrilateral shape of the captured video content;
to warp the cropped video content to an upright rectangle that adjusts the video content for perspective distortion due to a viewing angle of the camera relative to the display screen of the primary device, wherein a first perspective of the quadrilateral shape is different than a second perspective of the upright rectangle; and
to generate video fingerprints of the warped video content utilizing a fingerprint application running on the secondary device for transmission to a search function.
17. The system of claim 16 further comprising:
a searchable reference database storing reference fingerprints that match the generated fingerprints and utilized to identify the separated captured video content; and
memory storing matching reference fingerprints locally on the secondary device upon finding a match to the generated fingerprints.
18. The system of claim 17 further comprising:
the processing unit operating to search for the stored reference fingerprints on the secondary device to track the video content presented on the primary device.
19. The system of claim 16 further comprising:
a reference fingerprint database;
the processing unit operating to generate an action signal on the secondary device upon detection of a pre-specified video fragment stored in the reference fingerprint database; and
a software application running on the processing unit performing a pre-specified action based on local search results in response to the action signal.
20. The system of claim 16 , further comprising:
a display on the secondary device; and
a software application running on the processing unit presents information regarding local search results generated from the search function to a user of the secondary device.