Machine learning in video classification
Described herein are systems and methods that search videos and other media content to identify items, objects, faces, or other entities within the media content. Detectors identify objects within media content by, for instance, detecting a predetermined set of visual features corresponding to the objects. Detectors configured to identify an object can be trained using a machine learned model (e.g., a convolutional neural network) as applied to a set of example media content items that include the object. The systems provide user interfaces that allow users to review search results, pinpoint relevant portions of media content items where the identified objects are determined to be present, review detector performance and retrain detectors, providing search result feedback, and/or reviewing video monitoring results and analytics.
1. A computer-implemented method, comprising:
receiving, from a user, a search query;
classifying frames of a video using a set of detectors each configured to, for each frame of the video, process the frame and output a confidence score indicating a likelihood that the frame includes one of a set of objects associated with the search query is within the frame;
identifying sets of consecutive frames of the video, each set of consecutive frames associated with an object of the set of objects and associated with an above-threshold confidence score that the consecutive frames include the object; and
modifying a progress bar interface element to display highlighted locations within the progress bar interface element that correspond to each of the sets of consecutive frames of the video such that a highlighted location for each different object associated with the sets of consecutive frames comprises a different color;
wherein a set of consecutive frames corresponds to an average confidence score associated with a plurality of detectors that is greater than threshold confidence score.
2. The method of claim 1 , wherein the average confidence score comprises an average of a set of confidence scores associated with the plurality of detectors, and wherein at least one of the set of confidence scores is less than the threshold confidence score.
3. The method of claim 1 , wherein each set of consecutive frames comprises at least a threshold number of consecutive frames.
4. The method of claim 1 , wherein the threshold confidence score is based on the set of detectors.
5. The method of claim 1 , wherein the threshold confidence score is received from the user.
6. The method of claim 1 , wherein one or more confidence scores are displayed in conjunction with display of the video.
7. A non-transitory computer-readable storage medium storing executable instructions that, when executed by a hardware processor, cause the hardware processor to perform steps comprising:
receiving, from a user, a search query;
classifying frames of a video using a set of detectors each configured to, for each frame of the video, process the frame and output a confidence score indicating a likelihood that the frame includes one of a set of objects associated with the search query is within the frame;
identifying sets of consecutive frames of the video, each set of consecutive frames associated with an object of the set of objects and associated with an above-threshold confidence score that the consecutive frames include the object; and
modifying a progress bar interface element to display highlighted locations within the progress bar interface element that correspond to each of the sets of consecutive frames of the video such that a highlighted location for each different object associated with the sets of consecutive frames comprises a different color;
wherein a set of consecutive frames corresponds to an average confidence score associated with a plurality of detectors that is greater than threshold confidence score.
8. The non-transitory computer-readable storage medium of claim 7 , wherein the average confidence score comprises an average of a set of confidence scores associated with the plurality of detectors, and wherein at least one of the set of confidence scores is less than the threshold confidence score.
9. The non-transitory computer-readable storage medium of claim 7 , wherein each set of consecutive frames comprises at least a threshold number of consecutive frames.
10. The non-transitory computer-readable storage medium of claim 7 , wherein the threshold confidence score is based on the set of detectors.
11. The non-transitory computer-readable storage medium of claim 7 , wherein the threshold confidence score is received from the user.
12. The non-transitory computer-readable storage medium of claim 7 , wherein one or more confidence scores are displayed in conjunction with display of the video.
13. A system, comprising:
a computer processor; and
a non-transitory memory storing executable computer instructions that when executed by the computer processor are configured to cause the computer processor to perform steps comprising:
receiving, from a user, a search query;
classifying frames of a video using a set of detectors each configured to, for each frame of the video, process the frame and output a confidence score indicating a likelihood that the frame includes one of a set of objects associated with the search query is within the frame;
identifying sets of consecutive frames of the video, each set of consecutive frames associated with an object of the set of objects and associated with an above-threshold confidence score that the consecutive frames include the object; and
modifying a progress bar interface element to display highlighted locations within the progress bar interface element that correspond to each of the sets of consecutive frames of the video such that a highlighted location for each different object associated with the sets of consecutive frames comprises a different color;
wherein a set of consecutive frames corresponds to an average confidence score associated with a plurality of detectors that is greater than threshold confidence score.
14. The system of claim 13 , wherein the average confidence score comprises an average of a set of confidence scores associated with the plurality of detectors, and wherein at least one of the set of confidence scores is less than the threshold confidence score.
15. The system of claim 13 , wherein each set of consecutive frames comprises at least a threshold number of consecutive frames.
16. The system of claim 13 , wherein the threshold confidence score is based on the set of detectors.
17. The system of claim 13 , wherein the threshold confidence score is received from the user.