Identification model generation apparatus, identification apparatus, identification model generation method, identification method, and storage medium
An identification model generation apparatus comprises an acquisition unit configured to acquire the image of the eye and data on an image capture status of capturing the image of the eye, an inference unit configured to infer a feature amount of the eye based on the image of the eye and the data on the image capture status, and a learning unit configured to perform learning in the inference unit to approximate the feature amount of the eye obtained from the inference unit to ground truth data that is a ground truth value of the feature amount of the eye corresponding to the image of the eye and the data on the image capture status, and generate the identification model.
1. An identification model generation apparatus configured to generate an identification model for identifying a person using an image of an eye, the identification model generation apparatus comprising at least one processor or circuit configured to function as:
an acquisition unit configured to acquire the image of the eye and data on an image capture status of capturing the image of the eye,
an inference unit configured to infer a feature amount of the eye based on the image of the eye and the data on the image capture status, and
a learning unit configured to perform learning in the inference unit to approximate the feature amount of the eye obtained from the inference unit to ground truth data that is a ground truth value of the feature amount of the eye corresponding to the image of the eye and the data on the image capture status, and generate the identification model;
wherein the inference unit is a convolutional neural network configured to infer the feature amount of the eye responsive to an input of a combination of the image of the eye and the data on the image capture status.
2. The identification model generation apparatus according to claim 1 , wherein the feature amount of the eye is a feature amount of an iris of the eye.
3. The identification model generation apparatus according to claim 1 , wherein the data on the image capture status includes any one of image capturing time, image capturing location, a viewing direction of the eye, and a season.
4. The identification model generation apparatus according to claim 1 , wherein the learning unit calculates a weighting factor for the convolutional neural network to approximate the feature amount of the eye obtained from the inference unit to the ground truth data that is the ground truth value of the feature amount of the eye corresponding to the image of the eye and the data on the image capture status, based on the image of the eye and the data on the image capture status.
5. The identification model generation apparatus according to claim 1 , wherein the acquisition unit includes an image capturing device configured to capture an image of the eye.
6. The identification model generation apparatus according to claim 1 further comprising a detection device configured to detect brightness of the image of the eye.
7. The identification model generation apparatus according to claim 6 , wherein the detection device outputs, as the data on the image capture status, an evaluation value associating the brightness in the image of the eye with a position of each brightness.
8. The identification model generation apparatus according to claim 1 , wherein
the identification model generation apparatus further comprises a detection device configured to detect a predetermined characteristic of the image of the eye; and
the detection device outputs, as the data on the image capture status, an evaluation value associating the predetermined characteristic in the image of the eye with a position of each predetermined characteristic.
9. An identification apparatus configured to identify a person using an image of an eye, the identification apparatus comprising:
an image capturing device configured to capture the image of the eye; and
at least one processor or circuit configured to function as:
an acquisition unit configured to acquire the image of the eye captured by the image capturing device and data on an image capture status of capturing the image of the eye,
an inference unit configured to infer a feature amount of the eye based on the image of the eye and the data on the image capture status, and
a comparison unit configured to compare a feature amount of an eye of a specific person registered in advance with the feature amount of the eye inferred by the inference unit;
wherein the inference unit is a convolutional neural network configured to infer the feature amount of the eye responsive to an input of a combination of the image of the eye and the data on the image capture status.
10. The identification apparatus according to claim 9 , wherein the feature amount of the eye is a feature amount of an iris of the eye.
11. The identification apparatus according to claim 9 , wherein the data on the image capture status includes any one of image capturing time, image capturing location, a viewing direction of the eye, and a season.
12. The identification apparatus according to claim 9 further comprising a detection device configured to detect brightness of the image of the eye.
13. The identification apparatus according to claim 12 , wherein the detection device outputs, as the data on the image capture status, an evaluation value associating the brightness in the image of the eye with a position of each brightness.
14. An identification model generation method of generating an identification model for identifying a person using an image of an eye, the method comprising:
acquiring the image of the eye and data on an image capture status of capturing the image of the eye;
inferring a feature amount of the eye based on the image of the eye and the data on the image capture status; and
performing learning in the inferring to approximate the feature amount of the eye obtained by the inferring to ground truth data that is a ground truth value of the feature amount of the eye corresponding to the image of the eye and the data on the image capture status and generating the identification model;
wherein the inferring comprises inferring the feature amount of the eye by a convolutional neural network responsive to an input of a combination of the image of the eye and the data on the image capture status.
15. An identification method of identifying a person using an image of an eye, the method comprising:
capturing the image of the eye;
acquiring the image of the eye captured by the capturing and data on an image capture status of capturing the image of the eye;
inferring a feature amount of the eye based on the image of the eye and the data on the image capture status; and
comparing a feature amount of an eye of a specific person registered in advance with the feature amount of the eye inferred by the inferring;
wherein the inferring comprises inferring the feature amount of the eye by a convolutional neural network responsive to an input of a combination of the image of the eye and the data on the image capture status.
16. A non-transitory computer readable storage medium storing a program causing a computer to function as units of an identification model generation apparatus configured to generate an identification model for identifying a person using an image of an eye, the identification model generation apparatus comprising:
at least one processor or circuit configured to function as:
an acquisition unit configured to acquire the image of the eye and data on an image capture status of capturing the image of the eye,
an inference unit configured to infer a feature amount of the eye based on the image of the eye and the data on the image capture status, and
a learning unit configured to perform learning in the inference unit to approximate the feature amount of the eye obtained from the inference unit to ground truth data that is a ground truth value of the feature amount of the eye corresponding to the image of the eye and the data on the image capture status and generate the identification model;
wherein the inference unit is a convolutional neural network configured to infer the feature amount of the eye responsive to an input of a combination of the image of the eye and the data on the image capture status.
17. A non-transitory computer readable storage medium storing a program causing a computer to function as units of an identification apparatus configured to identify a person using an image of an eye, the identification apparatus comprising:
an image capturing device configured to capture the image of the eye; and
at least one processor or circuit configured to function as:
an acquisition unit configured to acquire the image of the eye captured by the image capturing device and data on an image capture status of capturing the image of the eye,
an inference unit configured to infer a feature amount of the eye based on the image of the eye and the data on the image capture status, and
a comparison unit configured to compare a feature amount of an eye of a certain person registered in advance with the feature amount of the eye inferred by the inference unit;
wherein the inference unit is a convolutional neural network configured to infer the feature amount of the eye responsive to an input of a combination of the image of the eye and the data on the image capture status.