Image processing device, image processing method, and recording medium in which program is stored

1. An image processing device comprising:

a memory storing a computer program; and

at least one processor configured to run the computer program to perform:

detecting, from time-series images, a first feature relating to a first motion of a head portion of a person and a second feature relating to a second motion of a body portion that is a region other than the head portion of the person;

calculating, as one integration feature, a weighted sum of the first feature relating to the first motion of the head portion of the person and the second feature relating to the second motion of the body portion of the person;

calculating, from the one integration feature, an index value indicating a degree of consistency between the first feature relating to the first motion of the head portion of the person and the second feature relating to the second motion of the body portion of the person; and

determining whether the person is spoofing another person, based on the index value, such that spoofing is detected based on a mismatch between the head portion and the body portion of the person.

2. The image processing device according to claim 1 , wherein

the index value is any one of:

a distance between a first motion vector indicating a direction in which the head portion of the person moves and a second motion vector indicating a direction in which the body portion of the person moves;

an angle between the first motion vector indicating the direction in which the head portion of the person moves and the second motion vector indicating the direction in which the body portion of the person moves; or

an output value of a function including a deep learning neural network, the function having the first motion vector indicating the direction in which the head portion of the person moves and the second motion vector indicating the direction in which the body portion of the person moves as inputs.

3. The image processing device according to claim 1 , wherein

the at least one processor is configured to run the computer program to perform:

detecting the head portion of the person from the time-series images;

detecting the region other than the head portion of the person from the time-series images;

extracting the first feature relating to the first motion of the head portion of the person from a detection result of the head portion of the person in the time-series images; and

extracting the second feature relating to the second motion of the body portion of the person from the detection result of the region other than the head portion of the person in the time-series images.

4. The image processing device according to claim 1 , wherein

the first feature relating to the first motion of the head portion of the person includes a first motion vector of the head portion of the person, and

the second feature relating to the second motion of the body portion of the person includes a second motion vector of the body portion of the person.

5. The image processing device according to claim 1 , wherein

the at least one processor is configured to run the computer program to perform:

detecting the first feature relating to the first motion of the head portion of the person and the second feature relating to the motion of the region other than the head portion of the person from the time-series images by using a neural network.

6. An image processing method comprising:

detecting, by a processor and from time-series images, a first feature relating to a first motion of a head portion of a person and a second feature relating to a second motion of a body portion that is a region other than the head portion of the person;

calculating, by the processor and as one integration feature, a weighted sum of the first feature relating to the first motion of the head portion of the person and the second feature relating to the second motion of the body portion of the person;

calculating, by the processor and from the one integration feature, an index value indicating a degree of consistency between the first feature relating to the first motion of the head portion of the person and the second feature relating to the second motion of the body portion of the person; and

determining, by the processor, whether the person is spoofing another person, based on the index value, such that spoofing is detected based on a mismatch between the head portion and the body portion of the person.

7. A non-transitory recording medium storing a program executable by a computer to perform:

detecting, from time-series images, a first feature relating to a first motion of a head portion of a person and a second feature relating to a second motion of a body portion that is a region other than the head portion of the person;

calculating, as one integration feature, a weighted sum of the first feature relating to the first motion of the head portion of the person and the second feature relating to the second motion of the body portion of the person;

calculating, from the one integration feature, an index value indicating a degree of consistency between the first feature relating to the first motion of the head portion of the person and the second feature relating to the second motion of the body portion of the person; and

determining whether the person is spoofing another person, based on the index value, such that spoofing is detected based on a mismatch between the head portion and the body portion of the person.

8. The image processing device according to claim 1 , wherein the index value when the first feature relating to the first motion of the head portion of the person and the second feature relating to the second motion of the body portion of the person does not match is lower than the index value when the first feature and the second feature match, and the at least one processor is configured to run the computer program to perform: determining that the lower the index value is, the higher a possibility that the person is spoofing.