System and method for person re-identification using overhead view images
A method, non-transitory computer readable medium and apparatus for performing a person re-identification using an overhead view image are disclosed. For example, the method includes receiving a plurality of overhead view images, detecting a target person in one or more of the plurality of overhead view images, creating a probe image of the target person, receiving a selection of the probe image containing the target person, selecting one or more of the plurality of processed images that has a similar distortion profile as a distortion profile of the probe image based on a radial distance of the target person from a center of a respective overhead view image of the plurality overhead view images used to generate the probe image and performing the person-re-identification of the target person using the one or more of the plurality of processed images that are selected.
1. A method for performing a person re-identification using an overhead view image, comprising:
receiving, by a processor, a plurality of overhead view images;
detecting, by the processor, a target person in one or more of the plurality of overhead view images;
creating, by the processor, a probe image of the target person;
receiving, by the processor, a selection of the probe image containing the target person;
selecting, by the processor, one or more of a plurality of processed images in an image gallery that has a similar distortion profile as a distortion profile of the probe image based on a radial distance of the target person from a center of a respective overhead view image of the plurality overhead view images used to generate the probe image; and
performing, by the processor, the person-re-identification of the target person in the probe image using the one or more of the plurality of processed images that are selected.
2. The method of claim 1 , wherein the plurality of overhead view images comprise images that are captured along a viewing axis that is substantially perpendicular to a ground.
3. The method of claim 1 , the plurality of overhead view images are captured by a fisheye camera.
4. The method of claim 1 , wherein the creating the plurality of processed images comprises:
detecting, by the processor, one or more persons in one or more of the plurality of overhead view images;
creating, by the processor, one or more additional images of each one of the one or more persons that are detected; and
normalizing, by the processor, the one or more additional images to create the plurality of processed images.
5. The method of claim 4 , wherein the one or more additional images are sub-images of a respective overhead view image of the plurality of overhead view images.
6. The method of claim 4 , wherein the normalizing comprises:
adjusting, by the processor, an orientation of a person in each one of the one or more additional images.
7. The method of claim 1 , wherein the plurality of processed images include different persons, wherein each one of the different persons is contained in a plurality of different processed images.
8. The method of claim 7 , wherein the selecting comprises:
selecting, by the processor, only a single one of the plurality of different processed images for the each one of the different persons.
9. The method of claim 7 , wherein the plurality of different processed images comprise different views.
10. The method of claim 9 , wherein the different views of the each one of the different persons are captured at different orientations at different radial distances from a center of a respective overhead view image of the plurality of overhead view images.
11. The method of claim 1 , wherein a person in each one of the one or more of the plurality of processed images that are selected is in a same orientation as the target person in the probe image.
12. A non-transitory computer-readable medium storing a plurality of instructions, which when executed by a processor, cause the processor to perform operations for performing a person re-identification using an overhead view image, the operations comprising:
receiving a plurality of overhead view images;
detecting a target person in one or more of the plurality of overhead view images;
creating a probe image of the target person;
receiving a selection of the probe image containing the target person;
selecting one or more of a plurality of processed images in an image gallery that has a similar distortion profile as a distortion profile of the probe image based on a radial distance of the target person from a center of a respective overhead view image of the plurality overhead view images used to generate the probe image; and
performing the person-re-identification of the target person in the probe image using the one or more of the plurality of processed images that are selected.
13. The non-transitory computer readable medium of claim 12 , wherein the plurality of overhead view images comprise images that are captured along a viewing axis that is substantially perpendicular to a ground.
14. The non-transitory computer readable medium of claim 12 , wherein the creating the plurality of processed images comprises:
detecting one or more persons in one or more of the plurality of overhead view images;
creating one or more additional images of each one of the one or more persons that are detected; and
normalizing the one or more additional images to create the plurality of processed images.
15. The non-transitory computer readable medium of claim 14 , wherein the one or more additional images are sub-images of a respective overhead view image of the plurality of overhead view images.
16. The non-transitory computer readable medium of claim 14 , wherein the normalizing comprises:
adjusting an orientation of a person in each one of the one or more additional images.
17. The non-transitory computer readable medium of claim 12 , wherein the plurality of processed images include different persons, wherein each one of the different persons is contained in a plurality of different processed images.
18. The non-transitory computer readable medium of claim 17 , wherein the selecting comprises:
selecting only a single one of the plurality of different processed images for the each one of the different persons.
19. The non-transitory computer readable medium of claim 17 , wherein the plurality of different processed images comprise different views, wherein the different views of the each one of the different persons are captured at different orientations at different radial distances from a center of a respective overhead view image of the plurality of overhead view images.
20. A method for performing a person re-identification using an overhead view image, comprising:
receiving, by a processor, a plurality of overhead view images;
detecting, by the processor, one or more persons in one or more of the plurality overhead view images;
creating, by the processor, a plurality of processed images of the one or more persons from the plurality of overhead view images, wherein the plurality of processed images are stored in a gallery of images;
detecting, by the processor, a target person in a second one or more of the plurality of overhead view images;
creating, by the processor, a probe image of the target person;
receiving, by the processor, a selection of the probe image containing the target person;
selecting, by the processor, one or more of the plurality of processed images that has a similar distortion profile as a distortion profile of the probe image based on a radial distance of the target person from a center of a respective overhead view image of the plurality overhead view images used to generate the probe image; and
performing, by the processor, the person-re-identification of the target person in the probe image using the one or more of the plurality of processed images that are selected.