Behavioral measurements in a video stream focalized on keywords

1. A computer-implemented method comprising:

extracting a plurality of salient fragments from a video, a salient fragment being a video sequence tracking a salient object through a subset of a series of frames in the video;

building a database of the plurality of salient fragments;

receiving a keyword;

identifying a time of utterance of the keyword in an audio track of the video;

retrieving one or more salient fragments of the video from the database of the plurality of salient fragments using the time of utterance of the keyword;

generating a video segment focalized on the utterance of the keyword using the one or more retrieved salient fragments of the video, the video segment starting at the time of utterance of the keyword;

tagging a human subject in the video segment with a unique identifier;

analyzing the video segment based on the time of utterance of the keyword and the unique identifier to generate a behavior score; and

providing the behavior score via a user device.

2. The method of claim 1 , further comprising:

storing the behavior score in a behavioral analysis database using the time of utterance of the keyword and the unique identifier as a key to the behavioral analysis database.

3. The method of claim 1 , wherein tagging the human subject in the video segment with the unique identifier comprises:

detecting a face in a frame of the video segment;

identifying the face using a template of a known human subject, wherein the template is associated with the unique identifier; and

associating the face with the unique identifier.

4. The method of claim 1 , further comprising:

tracking the human subject over a plurality of frames of the video segment.

5. The method of claim 1 , wherein analyzing the video segment based on the time of utterance of the keyword and the unique identifier to generate the behavior score comprises:

determining a position of the human subject in a fragment of the video segment at a time corresponding to the time of utterance of the keyword; and

assigning a score to the human subject as a function of the position of the human subject.

6. The method of claim 1 , wherein analyzing the video segment based on the time of utterance of the keyword and the unique identifier to generate the behavior score comprises:

determining a mood of the human subject in a fragment of the video segment at a time corresponding to the time of utterance of the keyword; and

assigning a score to the human subject as a function of the mood of the human subject.

7. The method of claim 1 , further comprising:

generating a baseline score for a behavioral attribute; and

determining a comparison score based on the baseline score and the behavior score.

8. A computer program product comprising a non-transitory computer readable medium storing a computer readable program, wherein the computer readable program when executed on a computer causes the computer to:

extract a plurality of salient fragments from a video, a salient fragment being a video sequence tracking a salient object through a subset of a series of frames in the video;

build a database of the plurality of salient fragments;

receive a keyword;

identify a time of utterance of the keyword in an audio track of the video;

retrieve one or more salient fragments of the video from the database of the plurality of salient fragments using the time of utterance of the keyword;

generate a video segment focalized on the utterance of the keyword using the one or more retrieved salient fragments of the video, the video segment starting at the time of utterance of the keyword;

tag a human subject in the video segment with a unique identifier;

analyze the video segment based on the time of utterance of the keyword and the unique identifier to generate a behavior score; and

provide the behavior score via a user device.

9. The computer program product of claim 8 , wherein the computer readable program when executed on the computer further causes the computer to:

store the behavior score in a behavioral analysis database using the time of utterance of the keyword and the unique identifier as a key to the behavioral analysis database.

10. The computer program product of claim 8 , wherein to tag the human subject in the video segment with the unique identifier, the computer readable program causes the computer to:

detect a face in a frame of the video segment;

identify the face using a template of a known human subject, wherein the template is associated with the unique identifier; and

associate the face with the unique identifier.

11. The computer program product of claim 8 , wherein the computer readable program when executed on the computer further causes the computer to:

track the human subject over a plurality of frames of the video segment.

12. The computer program product of claim 8 , wherein to analyze the video segment based on the time of utterance of the keyword and the unique identifier to generate the behavior score, the computer readable program causes the computer to:

determine a position of the human subject in a fragment of the video segment at a time corresponding to the time of utterance of the keyword; and

assign a score to the human subject as a function of the position of the human subject.

13. The computer program product of claim 8 , wherein to analyze the video segment based on the time of utterance of the keyword and the unique identifier to generate the behavior score, the computer readable program causes the computer to:

determine a mood of the human subject in a fragment of the video segment at a time corresponding to the time of utterance of the keyword; and

assign a score to the human subject as a function of the mood of the human subject.

14. The computer program product of claim 8 , wherein the computer readable program when executed on the computer further causes the computer to:

generate a baseline score for a behavioral attribute; and

determine a comparison score based on the baseline score and the behavior score.

15. A system comprising:

one or more processors; and

a memory, the memory storing instructions which when executed cause the one or more processors to:

extract a plurality of salient fragments from a video, a salient fragment being a video sequence tracking a salient object through a subset of a series of frames in the video;

build a database of the plurality of salient fragments;

receive a keyword;

identify a time of utterance of the keyword in an audio track of the video;

retrieve one or more salient fragments of the video from the database of the plurality of salient fragments using the time of utterance of the keyword;

generate a video segment focalized on the utterance of the keyword using the one or more retrieved salient fragments of the video, the video segment starting at the time of utterance of the keyword;

tag a human subject in the video segment with a unique identifier;

analyze the video segment based on the time of utterance of the keyword and the unique identifier to generate a behavior score; and

provide the behavior score via a user device.

16. The system of claim 15 , wherein the instructions further causes the one or more processors to:

store the behavior score in a behavioral analysis database using the time of utterance of the keyword and the unique identifier as a key to the behavioral analysis database.

17. The system of claim 15 , wherein to tag the human subject in the video segment with the unique identifier, the instructions further causes the one or more processors to:

detect a face in a frame of the video segment;

identify the face using a template of a known human subject, wherein the template is associated with the unique identifier; and

associate the face with the unique identifier.

18. The system of claim 15 , wherein the instructions further causes the one or more processors to:

track the human subject over a plurality of frames of the video segment.

19. The system of claim 15 , wherein to analyze the video segment based on the time of utterance of the keyword and the unique identifier to generate the behavior score, the instructions further causes the one or more processors to:

determine a position of the human subject in a fragment of the video segment at a time corresponding to the time of utterance of the keyword; and

assign a score to the human subject as a function of the position of the human subject.

20. The system of claim 15 , wherein to analyze the video segment based on the time of utterance of the keyword and the unique identifier to generate the behavior score, the instructions further causes the one or more processors to:

determine a mood of the human subject in a fragment of the video segment at a time corresponding to the time of utterance of the keyword; and

assign a score to the human subject as a function of the mood of the human subject.