Evaluation of models generated from objects in video
View Patent ↗Models are generated from objects identified in video. Each model is evaluated based on knowledge of the objects determined from video analysis, and preferred models are identified based on the evaluations. In some examples, each model could be evaluated by tracking a movement of each object in the video by using each model to track the object from which it was generated, evaluating an ability of each model to identify the objects in the video that are similar to the object from which it was generated, and determining an amount of false identifications made by each model of different objects in different video that does not include the object from which it was generated.
1. A method of operating an image processing system, the method comprising:
generating models from objects identified in video;
evaluating each model based on knowledge of the objects determined from video analysis;
identifying at least one preferred model based on the evaluating, wherein evaluating each model includes applying a set of video analytics in order to effectuate the identifying step, the evaluating further comprising:
selecting the model that detected the greatest number of objects in the video that are similar to the object from which it was generated, then removing those objects that it detected from the analysis;
selecting another model that detected the next greatest number of this same type of object in the video from among the remaining objects that were undetected by the first selected model; and so on;
tracking a movement of each object in the video, wherein tracking a movement of each object in the video comprises using a plurality of respective models to track the object from which it was generated, and
evaluating statistics based on the at least one preferred model,
wherein, the at least one preferred model identified avoids resemblance and redundancy among the models.
2. The method of claim 1 wherein evaluating each model based on knowledge of the objects determined from video analysis comprises tracking a movement of each object in the video.
3. The method of claim 1 wherein evaluating each model based on knowledge of the objects determined from video analysis comprises evaluating an ability of each model to identify the objects in the video that are similar to the object from which it was generated.
4. The method of claim 1 wherein evaluating each model based on knowledge of the objects determined from video analysis comprises determining an amount of false identifications made by each model of different objects in different video that does not include the object from which it was generated.
5. The method of claim 1 wherein evaluating each model based on knowledge of the objects determined from video analysis comprises:
tracking a movement of each object in the video by using each model to track the object from which it was generated;
evaluating an ability of each model to identify the objects in the video that are similar to the object from which it was generated; and
determining an amount of false identifications made by each model of different objects in different video that does not include the object from which it was generated.
6. The method of claim 5 wherein identifying at least one preferred model based on the evaluating comprises identifying a model having a greatest ability to identify the objects in the video that are similar to the object from which it was generated and having a least amount of false identifications of the different objects in the different video.
7. The method of claim 1 wherein the objects are identified in the video by manual identification.
8. The method of claim 1 wherein the objects are identified in the video by human head detection.
9. The method of claim 1 wherein the objects identified in the video comprise human body parts.
10. One or more computer readable media having stored thereon program instructions which, when executed by a processing system, direct the processing system to:
generate models from objects identified in video;
perform evaluations on each model based on knowledge of the objects determined from video analysis;
identify at least one preferred model based on the evaluations, wherein evaluating each model includes applying a set of video analytics in order to effectuate the identifying step, the evaluating further comprising selecting the model that detected the greatest number of objects in the video that are similar to the object from which it was generated, then removing those objects that it detected from the analysis; and selecting another model that detected the next greatest number of this same type of object in the video from among the remaining objects that were undetected by the first selected model; and so on;
tracking a movement of each object in the video, wherein tracking a movement of each object in the video comprises using a plurality of respective models to track the object from which it was generated; and
evaluate statistics based on the at least one preferred model, wherein, the at least one preferred model identified avoids resemblance and redundancy among the models.
11. The one or more computer readable media of claim 10 wherein the program instructions, to direct the processing system to perform the evaluations on each model based on knowledge of the objects determined from video analysis, direct the processing system to track a movement of each object in the video.
12. The one or more computer readable media of claim 10 wherein the program instructions, to direct the processing system to perform the evaluations on each model based on knowledge of the objects determined from video analysis, direct the processing system to evaluate an ability of each model to identify the objects in the video that are similar to the object from which it was generated.
13. The one or more computer readable media of claim 10 wherein the program instructions, to direct the processing system to perform the evaluations on each model based on knowledge of the objects determined from video analysis, direct the processing system to determine an amount of false identifications made by each model of different objects in different video that does not include the object from which it was generated.
14. The one or more computer readable media of claim 10 wherein the program instructions, to direct the processing system to perform the evaluations on each model based on knowledge of the objects determined from video analysis, direct the processing system to:
track a movement of each object in the video by using each model to track the object from which it was generated;
evaluate an ability of each model to identify the objects in the video that are similar to the object from which it was generated; and
determine an amount of false identifications made by each model of different objects in different video that does not include the object from which it was generated.
15. The one or more computer readable media of claim 14 wherein the program instructions, to direct the processing system to identify at least one preferred model based on the evaluations, direct the processing system to identify a model having a greatest ability to identify the objects in the video that are similar to the object from which it was generated and having a least amount of false identifications of the different objects in the different video.
16. The one or more computer readable media of claim 10 wherein the objects are identified in the video by manual identification.
17. The one or more computer readable media of claim 10 wherein the objects are identified in the video by human head detection.
18. An image processing system comprising:
a processing system configured to generate models from objects identified in video, perform evaluations on each model based on knowledge of the objects determined from video analysis, and identify at least one preferred model based on the evaluations, wherein evaluating each model includes applying a set of video analytics in order to effectuate the identifying step, the evaluating further comprising selecting the model that detected the greatest number of objects in the video that are similar to the object from which it was generated, then removing those objects that it detected from the analysis; and selecting another model that detected the next greatest number of this same type of object in the video from among the remaining objects that were undetected by the first selected model; and so on;
tracking a movement of each object in the video, wherein tracking a movement of each object in the video comprises using a plurality of respective models to track the object from which it was generated, and evaluate statistics based on the at least one preferred model, wherein, the at least one preferred model identified avoids resemblance and redundancy among the models.
19. The method of claim 1 , wherein an absence of at least one object is determined from the evaluated statistics based on the at least one preferred model.
20. The one or more computer readable media of claim 10 , wherein an absence of at least one object is determined from the evaluated statistics based on the at least one preferred model.