# US 2018/0322660 A1 · Machine learning classification and training for digital microscopy images

> Agents: this data is available as JSON from the patentgrep API and MCP server — see https://patentgrep.com/llms.txt

- Number: US 2018/0322660 A1
- Document ID: US20180322660A1
- Status: Published application
- Kind: A1
- Application number: 15/585,045
- Filed: 2017-05-02
- Published: 2018-11-08
- Earliest claimed: 2017-05-02
- Assignees: Techcyte, Inc.
- Applicants: TechCyte, Inc.
- Inventors: Richard Boyd Smith
- Web page: https://patentgrep.com/patent/US20180322660A1

## CPC

- G06V 10/7747: {Organisation of the process, e.g. bagging or boosting}
- G06F 18/2148: {characterised by the process organisation or structure, e.g. boosting cascade}
- G06F 18/24: Classification techniques
- G06T 7/97: {Determining parameters from multiple pictures (depth or shape recovery from multiple images G06T7/55; stereo camera calibration G06T7/85)}
- G06V 20/698: {Matching; Classification}
- G06T 2207/10056: Microscopic image
- G06T 2207/20081: Training; Learning
- G06V: IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06F: ELECTRIC DIGITAL DATA PROCESSING (computer systems based on specific computational models G06N)
- G06T: IMAGE DATA PROCESSING OR GENERATION, IN GENERAL

## Abstract

Systems, methods, and devices for training models or algorithms for classifying or detecting particles or materials in microscopy images are disclosed. A method includes receiving a plurality of microscopy images of a specimen and a classification for the specimen. The plurality of microscopy images includes a first image captured at a first magnification and a second image captured at the first magnification with a different focus than the first image. The method includes training a machine learning model or algorithm using the plurality of images, wherein the first image and the second image are provided with one or more labels indicating the classification.

## Claims

### Claim 1

A method comprising:
receiving a plurality of microscopy images of a specimen and a classification for the specimen, the plurality of microscopy images comprising a first image captured at a first magnification and a second image captured at the first magnification with a different focus than the first image; and
training a machine learning model or algorithm using the plurality of images, wherein the first image and the second image are provided with one or more labels indicating the classification such that the model can predict the label of an image with only one focal plane.

### Claim 2 (depends on 1)

The method of claim 1, wherein receiving the plurality of microscopy images of the specimen comprises capturing the plurality of images of the specimen, wherein:
capturing the first image comprises capturing at the first magnification with a first focal plane; and
capturing the second image comprises capturing at the first magnification with a second focal plane different than the first focal plane.

### Claim 3 (depends on 2)

The method of claim 2, wherein capturing the plurality of microscopy images of the specimen comprises capturing microscopy images of a sample by:
capturing a plurality of images with different focal planes of a first region; and
capturing a plurality of images with different focal planes of a second one or more regions, such that each captured region has a plurality of corresponding images with different focal planes.

### Claim 4 (depends on 3)

The method of claim 3, wherein each captured region has a plurality of corresponding images with different focal planes comprising two or more different focal planes.

### Claim 5 (depends on 1)

The method of claim 1, wherein training the machine learning model or algorithm comprises:
providing the first image with the classification as ground truth; and
providing the second image with the classification as ground truth.

### Claim 6 (depends on 1)

The method of claim 1, further comprising determining the classification, wherein determining the classification comprises determining a classification based on data not included in the plurality of microscopy images.

### Claim 7 (depends on 1)

The method of claim 1, further comprising determining the classification, wherein determining the classification comprises determining a classification based on a combination of data found in two or more of the plurality of microscopy images.

### Claim 8 (depends on 1)

The method of claim 1, wherein the machine learning model or algorithm comprises a deep neural network, wherein training the deep neural network comprises training using a back-propagation algorithm.

### Claim 9 (depends on 1)

The method of claim 1, wherein the specimen comprises a particle, wherein the classification comprises a classification for the particle, and wherein training comprises training the machine learning model or algorithm to determine a classification for a particle.

### Claim 10

A system for training a neural network, the system comprising:
one or more processors; and
non-transitory computer readable storage media storing instructions that, when executed by one or more processors, cause the one or more processors to:
receive a plurality of microscopy images of a specimen and a classification for the specimen, the plurality of microscopy images comprising a first image captured at a first magnification and a second image captured at the first magnification with a different focus than the first image; and
train a machine learning model or algorithm using the plurality of images, wherein the first image and the second image are provided with one or more labels indicating the classification.

### Claim 11 (depends on 10)

The system of claim 10, wherein receiving the plurality of microscopy images of the specimen comprises capturing the plurality of images of the specimen, wherein the instructions cause the one or more processors to control an imaging system to:
capture the first image at the first magnification with a first focal plane; and
capture the second image at the first magnification with a second focal plane different than the first focal plane.

### Claim 12 (depends on 11)

The system of claim 11, wherein capturing the plurality of microscopy images of the specimen comprises capturing microscopy images of a sample by:
capturing a plurality of images with different focal planes of a first region; and
capturing a plurality of images with different focal planes of a second one or more regions, such that each captured region has a plurality of corresponding images with different focal planes.

### Claim 13 (depends on 12)

The system of claim 12, wherein each captured region has a plurality of corresponding images with different focal planes comprising two or more different focal planes.

### Claim 14 (depends on 10)

The system of claim 10, wherein the instructions cause the one or more processors to train the machine learning model or algorithm by:
providing the first image with the classification as ground truth; and
providing the second image with the classification as ground truth.

### Claim 15 (depends on 10)

The system of claim 10, wherein the instructions cause the one or more processors to determine the classification, wherein determining the classification comprises obtaining classification information based on data not included in the plurality of microscopy images.

### Claim 16 (depends on 10)

The system of claim 10, wherein training the machine learning model or algorithm comprises training a deep neural network, wherein training the deep neural network comprises training using a back-propagation algorithm.

### Claim 17

Non-transitory computer readable storage media storing instructions that, when executed by one or more processors, cause the one or more processors to:
obtain microscopy images of a sample, wherein obtaining microscopy image comprises, for one or more regions of the sample:
capturing a first image with a first focal plane at a first magnification;
capturing one or more additional images with one or more additional focal planes at the first magnification;
training a neural network using the first image and the one or more additional images, wherein a label or ground truth for the first image and the one or more additional images is the same for a same region or same specimen captured in the first image and the one or more additional images.

### Claim 18 (depends on 17)

The computer readable storage media of claim 17, wherein the instructions cause the one or more processors to train the machine learning model or algorithm by training a deep neural network, wherein training the deep neural network comprises training using a back-propagation algorithm.

### Claim 19 (depends on 17)

The computer readable storage media of claim 17, wherein the instructions cause the one or more processors to determine the classification, wherein determining the classification comprises obtaining classification information based on data not included in the plurality of microscopy images.

### Claim 20 (depends on 17)

The computer readable storage media of claim 17, wherein each of the one or more regions has a plurality of corresponding images with different focal planes comprising two or more different focal planes.

### Claim 21

A system for training a neural network for mold spore classification, the system comprising:
means for receiving a plurality of microscopy images of a specimen and a classification for the specimen, the plurality of microscopy images comprising a first image captured at a first magnification and a second image captured at the first magnification with a different focus than the first image; and
means for training a machine learning model or algorithm using the plurality of images, wherein the first image and the second image are provided with one or more labels indicating the classification.
