Adaptive positioning of drones for enhanced face recognition
Presented herein are systems, methods and apparatuses for increasing reliability of face recognition in analysis of images captured by drone mounted imaging sensors, comprising: recognizing a target person in one or more iterations, each iteration comprising: identifying one or more positioning properties of the target person based on analysis of image(s) captured by imaging sensor(s) mounted on a drone operated to approach the target person, instructing the drone to adjust its position to an optimal facial image capturing position selected based on the positioning property(s), receiving facial image(s) of the target person captured by the imaging sensor(s), receiving a face classification associated with a probability score from machine learning model(s) trained to recognize the target person, and initiating another iteration in case the probability score does not exceed a certain threshold. Finally, the face classification may be outputted for use by one or more face recognition based systems.
1. A computer implemented method of increasing reliability of face recognition in analysis of images captured by drone mounted imaging sensors, comprising:
recognizing a target person in at least one iteration comprising:
identifying at least one positioning property of the target person based on analysis of at least one image captured by at least one imaging sensor mounted on a drone operated to approach the target person, the at least one imaging sensor locally controlled at the drone,
instructing the drone to adjust its position to an optimal facial image capturing position selected based on the at least one positioning property,
receiving at least one facial image of the target person captured by the at least one imaging sensor while the drone is located at the optimal facial image capturing position,
receiving a face classification associated with a probability score from at least one machine learning model trained to recognize the target person which is applied to the at least one facial image, wherein the at least one machine learning model is executed by at least one remote system connected to the drone via at least one network,
calculating an updated aggregated probability score by accumulating the probability score received in a current iteration of the at least one iteration to an aggregated score calculated before said current iteration such that in each iteration said aggregated probability score is increased by an amount of said probability score received in the current iteration, and
initiating another iteration in case the aggregated probability score does not exceed a certain threshold; and
outputting the face classification for use by at least one face recognition based system,
wherein the operation of the drone is controlled locally at the drone.
2. The computer implemented method of claim 1 , wherein the at least one positioning property is a member of a group consisting of: a head pose of the target person, a property of at least one potentially blocking object with respect to the target person and at least one environmental parameter affecting image capturing of the target person.
3. The computer implemented method of claim 2 , wherein the optimal facial image capturing position is defined by at least one position parameter of the drone, the at least one position parameter is a member of a group consisting of: a location of the drone with respect to the target person, a distance of the drone from the target person, an altitude of the drone with respect to the target person and a view angle of the at least one imaging sensor mounted on the drone with respect to the head pose identified for the target person.
4. The computer implemented method of claim 1 , further comprising adjusting at least one operational parameter of the at least one imaging sensor based on the at least one positioning property, the at least one operational parameter is a member of a group consisting of: a resolution, a zoom, a color, a field of view, an aperture, a shutter speed, a sensitivity (ISO), a white balance and an auto exposure.
5. The computer implemented method of claim 1 , wherein the at least one machine learning model is based on a neural network.
6. The computer implemented method of claim 1 , wherein at least part of the face classification is done by the at least one remote system connected to the drone via at least one network to receive the at least one image captured by the at least one imaging sensor mounted on the drone.
7. The computer implemented method of claim 1 , further comprising initiating a plurality of iterations to capture a plurality of facial images depicting the face of the target person from a plurality of view angles to form an at least partial three dimensional (3D) representation of at least part of a head of the target person.
8. The computer implemented method of claim 1 , wherein the face recognition based system is an automated delivery system using the face classification for authenticating an identity of the target person for delivery of goods to the target person.
9. The computer implemented method of claim 8 , further comprising initiating the at least one iteration for recognizing the target person in correlation with the goods at a time of delivery.
10. The computer implemented method of claim 9 , further comprising a location of the target person where the at least one image is captured for recognizing the target person is different form the location of the target person during the time of delivery.
11. The computer implemented method of claim 8 , further comprising instructing the drone to initiate at least one additional authentication sequence which is a member of a group consisting of: manual signature of the target person, a biometric authentication of the target person and a voice authentication of the target person.
12. A system for increasing reliability of face recognition in analysis of images captured by drone mounted imaging sensors, comprising:
at least one processor executing a code, the code comprising:
code instructions to recognize a target person in at least one iteration comprising:
identifying at least one positioning property of the target person based on analysis of at least one image captured by at least one imaging sensor mounted on a drone located in operated to approach the target person, the at least one imaging sensor locally controlled at the drone,
instructing the drone to adjust its position to an optimal facial image capturing position selected based on the at least one positioning property,
receiving at least one facial image of the target person captured by the at least one imaging sensor while the drone is located at the optimal facial image capturing position,
receiving a face classification associated with a probability score from at least one machine learning model trained to recognize the target person which is applied to the at least one facial image, wherein the at least one machine learning model is executed by at least one remote system connected to the drone via at least one network,
calculating an updated aggregated probability score by accumulating the probability score received in a current iteration of the at least one iteration to said aggregated score calculated before said current iteration such that in each iteration said aggregated probability score is increased by an amount of said probability score received in the current iteration, and
initiating another iteration in case the aggregated probability score does not exceed a certain threshold; and
code instructions to output the face classification for use by at least one face recognition based system,
wherein the operation of the drone is controlled locally at the drone.