Method of preparing a microscope for imaging a sample

Patent metadata

Field	Value
Publication number	US-12372770-B2
Application number	US-202017088479-A
Country	US
Kind code	B2
Filing date	Nov 3, 2020
Priority date	Nov 4, 2019
Publication date	Jul 29, 2025
Grant date	Jul 29, 2025

Abstract

A method for preparing a microscope for imaging a sample. The method includes the following steps: providing the microscope for imaging the sample, wherein the microscope comprises an objective having a motor-adjustable objective correction ring for correcting imaging aberrations; acquiring sample information comprising at least one of the following indications: thickness of a cover glass, material of the cover glass, sample temperature, sample type, sample location on a sample carrier, embedding medium of the sample, or immersion medium; imaging the sample using the microscope for generating at least one raw image of the sample with a first setting of the objective correction ring; and inputting the at least one raw image and the sample information into a machine algorithm and determining a second setting of the objective correction ring, which reduces the imaging aberrations vis à vis the first setting, by means of the algorithm on the basis of the raw image and the sample information.

First claim

The invention claimed is:

1. A method for preparing an optical microscope for optically imaging a sample, comprising the following steps:
a) providing the microscope for imaging the sample, wherein the microscope comprises an objective having an adjustable objective correction ring for correcting total imaging aberrations;
b) optically imaging the sample through the objective and onto an image sensor of the optical microscope for generating at least one raw image of the sample, wherein the objective correction ring is at a first setting value during the imaging;
c) inputting the at least one raw image of the sample into a machine-learned or machine learning algorithm;
d) evaluating the optical image quality of the at least one raw image of the sample by means of the algorithm to determine a second relative or absolute setting value of the objective correction ring, which second setting value differs from the first setting value and reduces the total imaging aberrations vis à vis the first setting, by means of the algorithm; and
e) adjusting the objective correction ring to the second setting value and providing the optical microscope for subsequent imaging of the sample through the objective and onto the image sensor.

2. The method as claimed in claim 1 , wherein the algorithm is embodied as a machine learning algorithm and carries out the following substeps in step d):
f1) determining a figure of merit for the raw image and adjusting the objective correction ring to a changed setting value, either in a simulation or during renewed imaging of the sample, and also iteratively carrying out the following steps f2)-f5):
f2) determining a further raw image of the sample for the changed setting of the objective correction ring;
f3) generating a current figure of merit for the raw image of the sample determined in step f2);
f4) checking whether the current figure of merit attained a minimum value, and ending the iteration using the last-determined setting value of the objective correction ring as the second setting if this is the case or a predefined number of iteration passes has been reached, otherwise continuing the iteration with step f5);
f5) adjusting the objective correction ring in the same direction as during the last adjustment if the current figure of merit is better than the previous one, otherwise adjusting the objective correction ring in the opposite direction, and continuing the iteration with step f2).

3. The method as claimed in claim 2 , wherein the figure of merit is based on an evaluation in accordance with a quality criterion comprising at least one of the following image properties: imaging sharpness; signal/noise ratio; presence, frequency or extent of predetermined artefacts; recognizability of predetermined structures.

4. The method as claimed in claim 1 , wherein the algorithm comprises a neural network, which assesses image quality and/or determines the second setting value.

5. The method as claimed in claim 1 , comprising acquiring sample information comprising at least one of the following indications: thickness of a cover glass, material of the cover glass, embedding medium of the sample, immersion medium, wherein in step d) the algorithm determines the second setting value on the basis of the sample information as well.

6. The method as claimed in claim 1 , wherein the algorithm determines at least one provisional setting value of the objective correction ring in step d) and steps b) and c) are iterated with the at least one provisional setting value.

7. The method as claimed in claim 1 , wherein in step d) the algorithm calculates a preview image from the at least one raw image and one the basis of the second setting value of the objective correction ring.

8. The method as claimed in claim 1 , wherein a point spread function (PSF) is provided for the microscope and step d) comprises to evaluate the point spread function (PSF).

9. The method as claimed in claim 1 , wherein the algorithm is carried out by a control device of the microscope and the objective correction ring is adjustable by motor and is set to the second setting by the control device.

10. A method for optically examining a sample by optical microscopy comprising the following steps:
a) providing a microscope comprising an objective having an adjustable objective correction ring for correcting total imaging aberrations:
b) optically imaging the sample through the objective and onto an image sensor of the optical microscope for generating at least one raw image of the sample wherein the objective correction ring is at a first setting value during the imaging;
c) inputting the at least one raw image of the sample into a machine-learned or machine learning algorithm;
d) evaluating the optical image quality of the at least one raw image of the sample by means of the algorithm to determine a second relative or absolute setting value of the objective correction ring, which second setting value differs from the first setting value and reduces the total imaging aberrations vis à vis the first setting value, by means of the algorithm on the basis of the raw image; and
e) subsequently imaging the sample to the image sensor through the objective while using the second setting value of the objective correction ring.

11. The method as claimed in claim 10 , wherein the algorithm is embodied as a machine learning algorithm and carries out the following substeps in step d):
f1) determining a figure of merit for the raw image and adjusting the objective correction ring to a changed setting value, either in a simulation or during renewed imaging of the sample, and also iteratively carrying out the following steps f2)-f5):
f2) determining a further raw image of the sample for the changed setting of the objective correction ring;
3) generating a current figure of merit for the raw image of the sample determined in step f2);
f4) checking whether the current figure of merit attained a minimum value, and ending the iteration using the last-determined setting value of the objective correction ring as second setting value if this is the case or a predefined number of iteration passes has been reached, otherwise continuing the iteration with step f5);
f5) adjusting the objective correction ring in the same direction as during the last adjustment if the current figure of merit is better than the previous one, otherwise adjusting the objective correction ring in the opposite direction, and continuing the iteration with step f2).

12. The method as claimed in claim 11 , wherein the figure of merit is based on an evaluation in accordance with a quality criterion comprising at least one of the following image properties: imaging sharpness; signal/noise ratio; presence, frequency or extent of predetermined artefacts; recognizability of predetermined structures.

13. The method as claimed in claim 10 , wherein the algorithm comprises a neural network, which assesses image quality and/or determines the second setting value.

14. The method as claimed in claim 10 , comprising acquiring sample information comprising at least one of the following indications: embedding medium of the sample, or immersion medium, wherein in step d) the algorithm determines the second setting value on the basis of the sample information as well.

15. The method as claimed in claim 10 , wherein the algorithm determines at least one provisional setting value of the objective correction ring in step d) and steps b) and c) are iterated with the at least one provisional setting value.

16. The method as claimed in claim 10 , wherein in step d) the algorithm c

Assignees

• Zeiss Carl Microscopy Gmbh

Inventors

• Amthor Manuel
• Haase Daniel
• Ohrt Thomas

Classifications

• G02B27/0025

  for optical correction, e.g. distorsion, aberration · CPC title

• G02B21/06

  Means for illuminating specimens · CPC title

• G06T2207/30168

  Image quality inspection · CPC title

• G06T2207/20084

  Artificial neural networks [ANN] · CPC title

• G06T2207/20081

  Training; Learning · CPC title