Mass image processing apparatus and method

Patent metadata

Field	Value
Publication number	US-12586169-B2
Application number	US-202318198440-A
Country	US
Kind code	B2
Filing date	May 17, 2023
Priority date	May 23, 2022
Publication date	Mar 24, 2026
Grant date	Mar 24, 2026

Abstract

A pre-processor applies a pre-process to an original mass image produced through mass spectrometry of a sample, to produce a model input image. An image quality converter has an image quality conversion model produced through machine learning based on a group of images produced by a scanning electron microscope, and produces a model output image through image quality conversion of the model input image. A post-processor applies a post-process to the model output image, to produce a mass image after image quality conversion.

First claim

The invention claimed is:

1 . A mass image processing apparatus comprising:
a producer configured to produce an original mass image of a sample based on a plurality of ionic strengths extracted from a plurality of mass spectra obtained from the sample; a pre-processor configured to apply a pre-process to the original mass image produced through mass spectrometry of the sample, to produce a model input image; a converter that has an image quality conversion model for improving an image quality of the original mass image of the sample, the image quality conversion model produced through machine learning based on a group of images produced through sample analysis different from the mass spectrometry, and the image quality conversion model configured to produce a model output image through image quality conversion of the model input image; and a post-processor configured to apply a post-process to the model output image, to produce a mass image after image quality conversion, wherein the pre-process is a process to fit the model input image with respect to an input condition of the converter, and the post-process is a process to fit the mass image after the image quality conversion with respect to a mass image output condition.

2 . The mass image processing apparatus according to claim 1 , wherein
the input condition of the converter comprises an intensity condition, and the pre-process comprises intensity scaling to fit an intensity distribution of the original mass image to the intensity condition.

3 . The mass image processing apparatus according to claim 2 , wherein
the pre-process further comprises a high intensity noise process to correct or remove a pixel value which satisfies a high intensity condition in the original mass image, and the pre-processor executes the intensity scaling after execution of the high intensity noise process.

4 . The mass image processing apparatus according to claim 3 , wherein
the pre-process further comprises a low intensity noise process to correct or remove a pixel value which satisfies a low intensity condition in the original mass image, and the pre-processor executes the intensity scaling after execution of the high intensity noise process and the low intensity noise process.

5 . The mass image processing apparatus according to claim 1 , wherein
the original mass image is an image produced through ion beam scanning or laser scanning with respect to the sample, the different sample analysis is sample observation by a scanning electron microscope, and the group of images are produced by the scanning electron microscope.

6 . A mass image processing apparatus comprising:
a pre-processor configured to apply a pre-process to an original mass image produced through mass spectrometry of a sample, to produce a model input image; a converter that has an image quality conversion model produced through machine learning based on a group of images produced through sample analysis different from the mass spectrometry, and is configured to produce a model output image through image quality conversion of the model input image; and a post-processor configured to apply a post-process to the model output image, to produce a mass image after image quality conversion, wherein the pre-process is a process to fit the model input image with respect to an input condition of the converter, and the post-process is a process to fit the mass image after the image quality conversion with respect to a mass image output condition, wherein the input condition of the converter comprises a first intensity condition and a first image size condition, the mass image output condition comprises a second intensity condition and a second image size condition, the pre-process comprises input-side intensity scaling to fit an intensity distribution of the original mass image with respect to the first intensity condition, and an input-side image manipulation to produce the model input image comprising an entirety or a part of the original mass image such that the first image size condition is satisfied, and the post-process comprises output-side intensity scaling to fit an intensity distribution of the model output image with respect to the second intensity condition, and an output-side image manipulation to produce the mass image after the image quality conversion including an entirety or a part of the model output image such that the second image size condition is satisfied.

7 . A method of processing a mass image, the method comprising:
producing an original mass image of a sample based on a plurality of ionic strengths extracted from a plurality of mass spectra obtained from the sample; applying a pre-process to the original mass image produced through mass spectrometry of the sample, to produce a model input image; producing a model output image from the model input image using an image quality conversion model for improving an image quality of the original mass image of the sample, the image quality conversion model produced through machine learning based on a group of images produced through sample analysis different from the mass spectrometry; and applying a post-process to the model output image, to produce a mass image after image quality conversion, wherein the pre-process is a process to fit the model input image with respect to an input condition of the image quality conversion model, and the post-process is a process to fit the mass image after the image quality conversion with respect to a mass image output condition.

8 . A non-transitory computer-readable storage medium storing a program which, when executed, causes an information processing apparatus to execute a mass image process, the mass image process comprising:
producing an original mass image of a sample based on a plurality of ionic strengths extracted from a plurality of mass spectra obtained from the sample; applying a pre-process to the original mass image produced through mass spectrometry of the sample, to produce a model input image; producing a model output image from the model input image using an image quality conversion model for improving an image quality of the original mass image of the sample, the image quality conversion model produced through machine learning based on a group of images produced through sample analysis different from the mass spectrometry; and applying a post-process to the model output image, to produce a mass image after image quality conversion, wherein the pre-process is a process to fit the model input image with respect to an input condition of the image quality conversion model, and the post-process is a process to fit the mass image after the image quality conversion with respect to a mass image output condition.

Assignees

• Jeol Ltd

Inventors

• Takei Masahiko
• Satoh Takaya

Classifications

• G06T5/77

  Retouching; Inpainting; Scratch removal · CPC title

• H01J49/26

  Mass spectrometers or separator tubes · CPC title

• G06T2207/10061

  from scanning electron microscope · CPC title

• G06T3/40

  Scaling of whole images or parts thereof, e.g. expanding or contracting · CPC title

• G06V10/60

  relating to illumination properties, e.g. using a reflectance or lighting model · CPC title