Image processing device and image processing system

1. An image processing device for calculating a feature amount of a person image from an image, the image processing device comprising:

a person image detector configured to detect a person image from an image;

a person feature amount extractor configured to calculate a person feature amount of the person image using an image recognition technique;

a person size calculator configured to calculate an image size of the person image; and

an extraction target selector configured to:

select a first person image having a largest image size from among a plurality of person images, detected by the person image detector, in a search person registration table as the first person image for extracting the person feature amount, and

select a second person image having at least a minimum cumulative number of times for extracting the person feature amount from among the plurality of person images in the search person registration table, excluding the first person image, as the second person image for extracting the person feature amount,

wherein the selecting the second person image having at least the minimum cumulative number of times for extracting includes:

selecting a person having a least number of times features have been extracted from the person in the search person registration table, excluding the first person image, having a current feature amount extraction flag of “0”,

increasing the number of times features have been extracted from the person by one, and

setting the current feature amount extraction flag of the person to “1”.

2. An image processing device for calculating a feature amount of a person image from an image, the image processing device comprising:

a person image detector configured to detect a person image from an image;

a person feature amount extractor configured to calculate a feature amount of the person image using an image recognition technique;

a person size calculator configured to calculate an image size of the person image;

a random number generator configured to generate a random number; and

an extraction target selector configured to:

select a first person image having a largest image size from among a plurality of person images, detected by the person image detector, in a search person registration table as the first person image for extracting the person feature amount, and

select a second person image by performing an arithmetic determination operation using the random number generated by the random number generator, from among the plurality of person images in the search person registration table, excluding the first person image, as the second person image for extracting the person feature amount,

wherein the arithmetic determination operation using the random number comprises:

determining a remainder of dividing the random number by a number of persons in the search person registration table, excluding the first person image, having a current feature amount extraction flag of “0”,

when the remainder is “0”, a person listed first in the search person registration table, excluding the first person image, is selected from among the persons having the current feature amount extraction flag of “0”, and

when the remainder is “1”, a person listed second in the search person registration table, excluding the first person image, is selected.

3. A method of processing an image, the method comprising:

receiving an image;

detecting a plurality of person images in the image;

identifying at least a first person image from among the plurality of person images in a search person registration table by selection processing;

identifying at least a second person image from among the plurality of person images, excluding the first person image, in the search person registration table by exhaustive processing; and

extracting a first feature amount as to the first person image and a second feature amount as to the second person image during a time interval,

wherein the exhaustive processing includes a determination processing, the determination processing comprising:

using a random number acquired in a random number acquisition processing;

determining a remainder of dividing the random number by a number of persons in the search person registration table, excluding the first person image, having a current feature amount extraction flag of “0”,

in response to the remainder of the dividing is “0”, a person listed first in the search person registration table, excluding the first person image, is selected from among persons having the current feature amount extraction flag of “0”, and

in response to the remainder of the dividing is “1”, a person listed second in the search person registration table, excluding the first person image, is selected.

4. The method of claim 3 , wherein a total number of a plurality of first person images and a plurality of second person images is no more than a maximum number of the plurality of person images based on processing capacity to processing the first and second plurality of person images during the time interval.

5. The method of claim 4 , wherein the selection processing includes determining a person image having a largest person size, among the detected plurality of person images in the image.

6. The method of claim 4 , wherein the selection processing includes determining a person image substantially facing a camera.

7. The method of claim 4 , wherein the selection processing includes determining a person image substantially being more likely to be a person by the detecting a plurality of person images in the image.

8. The method of claim 4 , wherein the exhaustive processing further includes determining a person image having a least number of times features have been extracted from the person in the search person registration table, excluding the first person image.

9. The method of claim 4 , wherein the exhaustive processing further includes determining a person image based at least in part on the random number.

10. The method of claim 5 , where the exhaustive processing further includes determining a person image having a least number of times features have been extracted from the person in the search person registration table, excluding the first person image.

11. The method of claim 10 , further comprising:

recording at least the first feature amount and the second feature amount.

12. The method of claim 5 , wherein the exhaustive processing further includes determining a person image based at least in part on the random number.

13. The method of claim 12 , further comprising:

detecting a similar person by referring to at least the first feature amount or the second feature amount.

14. The method of claim 3 , wherein the time interval includes a time interval consisting of a plurality of frames.

15. The method of claim 14 , wherein a total number of the plurality of first person images and the plurality of second person images is no more than a maximum number of the plurality of person images based on processing capacity to processing the first and second plurality of person images during the time interval consisting of the plurality of frames.

16. The method of claim 15 , wherein identifying the first person image by the selection processing is performed by referring to at least a frame, received during the time interval consisting of the plurality of frames, and identifying the second person image by the exhaustive processing is performed by referring to at least another frame, received during the time interval consisting of the plurality of frames.

17. The method of claim 16 , wherein the selection processing includes determining a person image having at least a largest person size, among the detected plurality of person images in the image.

18. The method of claim 16 , wherein the exhaustive processing further includes determining a person image having a least number of times features have been extracted from the person, excluding the first person image.

19. The method of claim 16 , wherein the exhaustive method further includes determining a person image based at least in part on a random number.

20. The method of claim 15 , wherein identifying the first person image by the selection processing is performed at either an odd frame or an even frame, and identifying the second person image by the exhaustive processing is performed at a frame having another parity.