Techniques for indexing video files
In one embodiment, an apparatus may include a processor circuit and a keyframe indexing module that is operative on the processor circuit to map sensor data from a first sensor to a set of video data, where the sensor data and video data correspond to an event. The keyframe indexing module may be further operative on the processor circuit to identify one or more key portions of the set of sensor data, index the one or more key portions to respective mapped one or more video frames of the video data, and generate a first keyframe index to identify the one or more video frames.
1. An apparatus, comprising:
a processor circuit; and
a keyframe indexing module operative on the processor circuit to:
map first non-audio sensor data from a first sensor to video data, the first non-audio sensor data and video data corresponding to an event;
identify one or more first key portions of the first non-audio sensor data based on a change in one or more successive data values of the first non-audio sensor data that exceeds a first threshold change value, to identify the one or more first key portions includes:
monitoring first non-audio sensor data as a function of time; and
assigning a first key portion to first non-audio sensor data within an interval in which a value of first non-audio sensor data vs. time or rate of change of the value of first non-audio sensor data vs. time deviates from another value of first non-audio sensor data vs. time or rate of change of the other first non-audio sensor data vs. time in an immediately prior interval by the first threshold change value;
index the one or more first key portions to respective mapped one or more first video frames of the video data;
generate a first keyframe index to identify the one or more first video frames;
map second non-audio sensor data from a second sensor to the video data, the second non-audio sensor data corresponding to the event;
identify one or more second key portions of the second non-audio sensor data based on a change in one or more successive data values of the second non-audio sensor data that exceeds a second threshold change value, to identify the one or more second key portions includes:
monitoring second non-audio sensor data as a function of time; and
assigning a second key portion to second non-audio sensor data within an interval in which a value of second non-audio sensor data vs. time or rate of change of the value of second non-audio sensor data vs. time deviates from another value of second non-audio sensor data vs. time or rate of change of the other second non-audio sensor data vs. time in an immediately prior interval by the second threshold change value;
index the one or more second key portions of the second non-sensor data to respective mapped one or more second video frames of the video data; and
generate a second keyframe index to identify the one or more second video frames.
2. The apparatus of claim 1 , the first or second non-audio sensor data comprising data recorded by an accelerometer, a gyrometer, a pressure sensor, a temperature sensor, a light sensor, or a humidity sensor.
3. The apparatus of claim 1 , the keyframe indexing module operative on the processor circuit to:
map audio data to the video data, the audio data and video data captured during recording of the event;
identify key portions of the audio data; and
generate a third keyframe index identifying one or more video frames of the video data based on indexing the set of key portions of the audio data to respective mapped video frames of the video data.
4. The apparatus of claim 3 , the first or the second keyframe index and the third keyframe index comprising the same keyframe index.
5. The apparatus of claim 1 , the keyframe indexing module operative on the processor circuit to:
generate an updated first or second keyframe index by storing a set of manually entered changes to the respective first or second keyframe indexes; and
modify a procedure to identify third key portions of third non-audio sensor data captured during recording of a second event according to the manually entered changes.
6. The apparatus of claim 5 , the keyframe indexing module operative on the processor circuit to:
identify missing keyframes of the first or the second keyframe index, the missing keyframes comprising keyframes of the first or the second keyframe index that are not present in the updated first or updated second keyframe index; and
assign the missing keyframes as false-positive classification errors.
7. The apparatus of claim 5 , the keyframe indexing module operative on the processor circuit to:
identify new keyframes of the first or the second keyframe index, the new keyframes comprising keyframes of the updated first or updated second keyframe index that are not present in the first or the second keyframe index; and
assign the new keyframes as false-negative classification errors.
8. The apparatus of claim 1 , the keyframe indexing module operative on the processor circuit to identify the one or more first or second key portions of the first or second non-audio sensor data by:
assigning a respective first or second key portion to first or second non-audio sensor data within an interval in which a value of first or second non-audio sensor data vs. time changes sign.
9. At least one non-transitory computer-readable storage medium comprising instructions that, when executed, cause a system to:
map first non-audio sensor data from a first sensor to video data, the first non-audio sensor data and video data corresponding to an event;
identify one or more first key portions of the first non-audio sensor data based on a change in one or more successive data values of the first non-audio sensor data that exceeds a first threshold change value, to identify the one or more first key portions includes:
monitoring first non-audio sensor data as a function of time; and
assigning a first key portion to first non-audio sensor data within an interval in which a value of first non-audio sensor data vs. time or rate of change of the value of first non-audio sensor data vs. time deviates from another value of first non-audio sensor data vs. time or rate of change of the other first non-audio sensor data vs. time in an immediately prior interval by the first threshold change value;
index the one or more first key portions to respective mapped one or more first video frames of the video data;
generate a first keyframe index to identify the one or more first video frames;
map second non-audio sensor data from a second sensor to the video data, the second non-audio sensor data corresponding to the event;
identify one or more second key portions of the second non-audio sensor data based on a change in one or more successive data values of the second non-audio sensor data that exceeds a second threshold change value, to identify the one or more second key portions includes:
monitoring second non-audio sensor data as a function of time; and
assigning a second key portion to second non-audio sensor data within an interval in which a value of second non-audio sensor data vs. time or rate of change of the value of second non-audio sensor data vs. time deviates from another value of second non-audio sensor data vs. time or rate of change of the other second non-audio sensor data vs. time in an immediately prior interval by the second threshold change value;
index the one or more second key portions of the second non-sensor data to respective mapped one or more second video frames of the video data; and
generate a second keyframe index to identify the one or more second video frames.
10. The at least one non-transitory computer-readable storage medium of claim 9 , the first or second non-audio sensor data comprising data recorded by an accelerometer, a gyrometer, a pressure sensor, a temperature sensor, a light sensor, or a humidity sensor.
11. The at least one non-transitory computer-readable storage medium of claim 9 comprising instructions that, when executed, cause a system to:
map audio data to the video data, the audio data and video data captured during recording of the event;
identify key portions of the audio data; and
generate a third keyframe index identifying one or more video frames of the video data based on indexing the set of key portions of the audio data to respective mapped video frames of the video data.
12. The at least one non-transitory computer-readable storage medium of claim 11 , the first or the second keyframe index and the third keyframe index comprising the same keyframe index.
13. The at least one non-transitory computer-readable storage medium of claim 9 comprising instructions that, when executed, cause a system to:
generate an updated first or second keyframe index by storing a set of manually entered changes to the respective first or second keyframe indexes; and
modify a procedure to identify third key portions of third non-audio sensor data captured during recording of a second event according to the manually entered changes.
14. The at least one non-transitory computer-readable storage medium of claim 13 comprising instructions that, when executed, cause a system to:
identify missing keyframes of the first or the second keyframe index, the missing keyframes comprising keyframes of the first or the second keyframe index that are not present in the updated first or updated second keyframe index; and
assign the missing keyframes as false-positive classification errors.
15. The at least one non-transitory computer-readable storage medium of claim 13 comprising instructions that, when executed, cause a system to:
identify new keyframes of the first or the second keyframe index, the new keyframes comprising keyframes of the updated first or updated second keyframe index that are not present in the first or the second keyframe index; and
assign the new keyframes as false-negative classification errors.
16. The at least one non-transitory computer-readable storage medium of claim 9 comprising instructions that, when executed, cause a system to identify the one or more first or second key portions of the first or second non-audio sensor data by:
assigning a respective first or second key portion to first or second non-audio sensor data within an interval in which a value of first or second non-audio sensor data vs. time changes sign.
17. A computer implemented method, comprising:
mapping first non-audio sensor data from a first sensor to video data, the first non-audio sensor data and video data corresponding to an event;
identifying one or more first key portions of the first non-audio sensor data based on a change in one or more successive data values of the first non-audio sensor data that exceeds a first threshold change value, identifying the one or more first key portions includes:
monitoring first non-audio sensor data as a function of time; and
assigning a first key portion to first non-audio sensor data within an interval in which a value of first non-audio sensor data vs. time or rate of change of the value of first non-audio sensor data vs. time deviates from another value of first non-audio sensor data vs. time or rate of change of the other first non-audio sensor data vs. time in an immediately prior interval by the first threshold change value;
indexing the one or more first key portions to respective mapped one or more first video frames of the video data;
generating a first keyframe index to identify the one or more first video frames;
mapping second non-audio sensor data from a second sensor to the video data, the second non-audio sensor data corresponding to the event;
identifying one or more second key portions of the second non-audio sensor data based on a change in one or more successive data values of the second non-audio sensor data that exceeds a second threshold change value, to identify the one or more second key portions includes:
monitoring second non-audio sensor data as a function of time; and
assigning a second key portion to second non-audio sensor data within an interval in which a value of second non-audio sensor data vs. time or rate of change of the value of second non-audio sensor data vs. time deviates from another value of second non-audio sensor data vs. time or rate of change of the other second non-audio sensor data vs. time in an immediately prior interval by the second threshold change value;
indexing the one or more second key portions of the second non-sensor data to respective mapped one or more second video frames of the video data; and
generating a second keyframe index to identify the one or more second video frames.
18. The computer implemented method of claim 17 , the first or second non-audio sensor data comprising data recorded by an accelerometer, a gyrometer, a pressure sensor, a temperature sensor, a light sensor, or a humidity sensor.
19. The computer implemented method of claim 17 , comprising:
mapping audio data to the video data, the audio data and video data captured during recording of the event;
identifying key portions of the audio data; and
generating a third keyframe index identifying one or more video frames of the video data based on indexing the set of key portions of the audio data to respective mapped video frames of the video data.
20. The computer implemented method of claim 19 , the first or the second keyframe index and the third keyframe index comprising the same keyframe index.
21. The computer implemented method of claim 17 , comprising:
generating an updated first or second keyframe index by storing a set of manually entered changes to the respective first or second keyframe indexes; and
modifying a procedure to identify third key portions of third non-audio sensor data captured during recording of a second event according to the manually entered changes.
22. The computer implemented method of claim 21 , comprising:
identifying missing keyframes of the first or the second keyframe index, the missing keyframes comprising keyframes of the first or the second keyframe index that are not present in the updated first or updated second keyframe index; and
assigning the missing keyframes as false-positive classification errors.
23. The computer implemented method of claim 21 , comprising:
identifying new keyframes of the first or the second keyframe index, the new keyframes comprising keyframes of the updated first or updated second keyframe index that are not present in the first or the second keyframe index; and
assigning the new keyframes as false-negative classification errors.
24. The computer implemented method of claim 17 , comprising identifying the one or more first or second key portions of the first or second non-audio sensor data by:
assigning a respective first or second key portion to first or second non-audio sensor data within an interval in which a value of first or second non-audio sensor data vs. time changes sign.