Image processing apparatus, image processing method, and image processing program

1. An image processing apparatus comprising:

a memory, and

a processor coupled to the memory and configured to execute an image recognition process comprising:

changing a rendering parameter in a three-dimensional model of a recognition target, the rendering parameter includes at least one or any combination of ones selected from texture for the three-dimensional model of the recognition target, a shape of the three-dimensional model of the recognition target, a camera position for the three-dimensional model of the recognition target, a camera orientation for the three-dimensional model of the recognition target, a camera viewing angle for the three-dimensional model of the recognition target, an illumination condition of the three-dimensional model of the recognition target, a background three-dimensional model, a masking ratio, and a storage condition of a teacher image;

generating teacher data of the recognition target based on the rendering parameter changed by the changing; and

generating a teacher image having the masking ratio that is equal to or smaller than a masking ratio threshold, wherein the masking ratio threshold is set for each class of the three-dimensional model of the recognition target.

2. The image processing apparatus according to claim 1 , the process further comprising: generating a camera model by changing at least one of the camera position, the camera orientation or the camera viewing angle for the three-dimensional model of the recognition target.

3. The image processing apparatus according to claim 1 , the process further comprising: generating an illumination model by changing the illumination condition for the three-dimensional model of the recognition target.

4. The image processing apparatus according to claim 1 , the process further comprising: fetching a background three-dimensional model from a storage unit.

5. The image processing apparatus according to claim 1 , the process further comprising: performing a machine learning using the generated teacher data.

6. The image processing apparatus according to claim 5 , the process further comprising: performing an inference using a learned weight found in the performing.

7. An image processing method performed by a computer, the method comprising:

changing a rendering parameter in a three-dimensional model of a recognition target, the rendering parameter includes at least one or any combination of ones selected from texture for the three-dimensional model of the recognition target, a shape of the three-dimensional model of the recognition target, a camera position for the three-dimensional model of the recognition target, a camera orientation for the three-dimensional model of the recognition target, a camera viewing angle for the three-dimensional model of the recognition target, an illumination condition of the three-dimensional model of the recognition target, a background three-dimensional model, a masking ratio, and a storage condition of a teacher image;

generating teacher data of the recognition target based on the rendering parameter changed by the changing; and

generating a teacher image having the masking ratio that is equal to or smaller than a masking ratio threshold, wherein the masking ratio threshold is set for each class of the three-dimensional model of the recognition target.

8. A non-transitory computer-readable medium storing an image processing program for causing a computer to perform an image recognition process, the process comprising:

changing a rendering parameter in a three-dimensional model of a recognition target, the rendering parameter includes at least one or any combination of ones selected from texture for the three-dimensional model of the recognition target, a shape of the three-dimensional model of the recognition target, a camera position for the three-dimensional model of the recognition target, a camera orientation for the three-dimensional model of the recognition target, a camera viewing angle for the three-dimensional model of the recognition target, an illumination condition of the three-dimensional model of the recognition target, a background three-dimensional model, a masking ratio, and a storage condition of a teacher image;

generating teacher data of the recognition target based on the rendering parameter changed by the changing; and

generating a teacher image having the masking ratio that is equal to or smaller than a masking ratio threshold, wherein the masking ratio threshold is set for each class of the three-dimensional model of the recognition target.

9. An image processing apparatus comprising:

a memory storing instructions;

a central processor coupled to the memory; and

a graphics processor coupled to the memory and the central processor, the instructions causing the central processor and the graphics processor to execute an image recognition process of:

acquiring a recognition target;

rendering, by the graphics processor, a three-dimensional model of the recognition target;

changing a parameter of the three-dimensional model, the rendering parameter includes at least one or any combination of ones selected from texture for the three-dimensional model of the recognition target, a shape of the three-dimensional model of the recognition target, a camera position for the three-dimensional model of the recognition target, a camera orientation for the three-dimensional model of the recognition target, a camera viewing angle for the three-dimensional model of the recognition target, an illumination condition of the three-dimensional model of the recognition target, a background three-dimensional model, a masking ratio, and a storage condition of a teacher image;

generating, by the central processor, teacher data of the recognition target based on the parameter changed; and

generating a teacher image having the masking ratio that is equal to or smaller than a masking ratio threshold, wherein the masking ratio threshold is set for each class of the three-dimensional model of the recognition target.