Microscope system
An acquisition condition is decided for a second image of improved quality. Values x i resulting from down sampling brightness values of an input first image are accepted by an input layer. A filter is scanned and a convolutional computation performed in a convolutional layer. Outputs z 1 to z 4 of the convolutional layer and a first image acquisition condition v=(v 1 , v 2 , v 3 , v 4 ) of the first image are accepted by an output layer and a second acquisition condition y is computed by the output layer.
1. A microscope system comprising:
a storage portion that stores a model of computation, which receives one or more input signals to decide one or more output signals; and
a processor that, using the model of computation, determines an acquisition condition of a second image based on (i) a first image and (ii) an acquisition condition of the first image,
wherein the model of computation corresponds to a type of subject for acquiring an image, and
wherein the processor:
determines the second image acquisition condition using the model of computation corresponding to the type of subject;
updates the model of computation using teaching data, the teaching data including (i) a first teaching image and an acquisition condition of the first teaching image, (ii) a second teaching image and an acquisition condition of the second teaching image, and (iii) a setting value of an acquisition condition corresponding to an image with maximum image quality;
generates the first and second teaching images and selects the second teaching image from a plurality of images; and
decides the second teaching image based on a score representing a quality of each of the plurality of images.
2. The microscope system of claim 1 , wherein the teaching data with which the processor updates the model of computation corresponds to the type of subject.
3. The microscope system of claim 1 , wherein the setting value corresponding to the image with maximum image quality is included in an acquisition condition under which the second teaching image was acquired.
4. The microscope system of claim 1 , wherein:
the first image is used as the first teaching image;
the acquisition condition of the first image is used as the acquisition condition of the first teaching image;
the second image is used as the second teaching image; and
the acquisition condition of the second image is used as the acquisition condition of the second teaching image.
5. The microscope system of claim 1 , wherein the acquisition condition of the second image includes a plurality of items.
6. The microscope system of claim 1 , wherein the model of computation is generated using machine learning.
7. The microscope system of claim 1 , wherein the model of computation includes a neural network.
8. The microscope system of claim 1 , further comprising:
an illumination optical system configured to illuminate an object with light emitted from a light source;
a detection section configured to detect light from the object; and
an image generation section configured to generate an image by employing the detected light.
9. A microscope system comprising:
a storage portion that stores a model of computation, which receives one or more input signals to determine one or more output signals; and
a processor (1) that, using the model of computation, determines an acquisition condition of a second image based on (i) a first image and (ii) an acquisition condition of the first image and (2) that updates the model of computation using teaching data, wherein:
the teaching data includes a first teaching image and an acquisition condition of the first teaching image, a second teaching image and an acquisition condition of the second teaching image, and a setting value of an acquisition condition corresponding to an image with maximum image quality;
the processor generates the first and second teaching images and selects the second teaching image from a plurality of images; and
the processor decides the second teaching image based on a score representing a quality of each of the plurality of images.
10. The microscope system of claim 9 , wherein the setting value corresponding to the image with maximum image quality is included in an acquisition condition under which the second teaching image was acquired.
11. The microscope system of claim 9 , wherein:
the first image is used as the first teaching image;
the acquisition condition of the first image is used as the acquisition condition of the first teaching image;
the second image is used as the second teaching image; and
the acquisition condition of the second image is used as the acquisition condition of the second teaching image.
12. The microscope system of claim 9 , wherein the acquisition condition of the second image includes a plurality of items.
13. The microscope system of claim 9 , wherein the model of computation is generated using machine learning.
14. The microscope system of claim 9 , wherein the model of computation includes a neural network.
15. The microscope system of claim 9 , comprising:
an illumination optical system configured to illuminate an object with light emitted from a light source;
a detection section configured to detect light from the object; and
an image generation section configured to generate an image by employing the detected light.