Processing recipe generation device

1. A processing recipe generation device that generates a processing recipe, the processing recipe describing an operating condition of the processing device when the processing devices processes an object, the processing recipe generation device comprising:

a storage unit configured to store the processing recipe of the processing device, an inspection image, and a correspondence relation between the processing recipe and the inspection image, the object processed by the processing device using the processing recipe being captured in the inspection image;

an inspection image conversion unit configured to generate a converted image, the converted image being generated by converting the inspection image using a conversion condition to reduce components relying on an imaging condition of the inspection image among components that the inspection image has;

a learning unit configured to learn the relation between the processing recipe and the converted image;

a target shape input unit configured to input a target shape processed by the processing device;

a target image generation unit configured to generate a target image, the target image being generated by converting an image representing the target shape in accordance with the conversion condition using a learning result by the learning unit; and

a recipe generation unit configured to generate a target processing recipe used to process a shape represented by the target image.

2. The processing recipe generation device according to claim 1 , wherein

the inspection image conversion unit is configured to divide the inspection image into one or more sub regions, and the inspection image conversion unit is configured to provide pixel values different depending on each of the sub regions to generate the converted image.

3. The processing recipe generation device according to claim 1 , wherein

the target image generation unit is configured to generate the target image divided into one or more sub regions and having pixel values different depending on each of the sub regions.

4. The processing recipe generation device according to claim 1 , wherein

the learning unit includes a first predictor, and the first predictor is configured to receive the processing recipe as an input and predict a predicted conversion image corresponding to the input processing recipe based on the result of learning the relation between the processing recipe and the converted image.

5. The processing recipe generation device according to claim 4 , wherein

the learning unit includes:

a second predictor configured to receive the converted image as an input, the second predictor being configured to predict a predicted processing recipe corresponding to the input converted image based on the result of learning the relation between the processing recipe and the converted image; and

an identifier configured to receive the processing recipe or the predicted processing recipe as an input and receive the converted image or the predicted converted image as an input, the identifier being configured to identify whether the input recipe and the input image are in a correct combination, and

the learning unit is configured to cause the first predictor to learn so as to output the predicted conversion image by which the identifier outputs an incorrect identification result, the learning unit is configured to cause the second predictor to learn so as to output the predicted processing recipe by which the identifier outputs an incorrect identification result, and the learning unit is configured to repeatedly cause the identifier to learn such that the identifier outputs a correct identification result to learn the relation between the processing recipe and the conversion image.

6. The processing recipe generation device according to claim 5 , wherein

the first predictor and the second predictor are configured to perform learning such that a result of prediction by receiving the predicted converted image as an input by the second predictor matches the processing recipe, and

the first predictor and the second predictor are configured to perform learning such that a result of prediction by receiving the predicted processing recipe as an input by the first predictor matches the converted image.

7. The processing recipe generation device according to claim 4 , wherein

the recipe generation unit is configured to search for a processing recipe with which a difference between the predicted converted image and the target image becomes small as much as possible from a processing recipe candidate group to generate the target processing recipe.

8. The processing recipe generation device according to claim 7 , wherein

the recipe generation unit is configured to employ a processing recipe with which the difference becomes a local minimum or a processing recipe with which a changing rate of the difference becomes a local minimum as the target processing recipe.

9. The processing recipe generation device according to claim 1 , further comprising

a condition setting unit configured to designate a conversion condition when the inspection image conversion unit converts the inspection image.

10. The processing recipe generation device according to claim 9 , wherein

the condition setting unit is configured to designate, as the conversion conditions:

an imaging magnification of the converted image;

an image size of the converted image; and

a reference point of positioning between the inspection image and the converted image, and

the target image generation unit is configured to generate the target image such that the imaging magnification of the target image, the image size of the target image, and the reference point in the target image match the imaging magnification of the converted image, the image size of the converted image, and the reference point of the positioning between the inspection image and the converted image designated by the condition setting unit, respectively.