# US 10,769,432 B2 · Automated parameterization image pattern recognition method

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

- Number: US 10,769,432 B2
- Document ID: US10769432B2
- Status: Granted
- Kind: B2
- Application number: 16/156,814
- Filed: 2018-10-10
- Published: 2020-04-16
- Granted: 2020-09-08
- Earliest claimed: 2018-10-10
- Assignees: DRVision Technologies LLC
- Applicants: DRVision Technologies LLC
- Inventors: Michael William Jones; Luciano Andre Guerreiro Lucas; Hoyin Lai; Casey James McBride; Shih-Jong James Lee
- Web page: https://patentgrep.com/patent/US10769432B2

## CPC

- G06N 20/00: Machine learning
- G06T 7/11: Region-based segmentation
- G06T 7/155: involving morphological operators
- G06T 7/187: involving region growing; involving region merging; involving connected component labelling
- G06V 10/34: Smoothing or thinning of the pattern; Morphological operations; Skeletonisation
- G06V 10/764: using classification, e.g. of video objects
- G06V 20/695: {Preprocessing, e.g. image segmentation}
- G06F 2218/08: Feature extraction
- G06F 2218/12: Classification; Matching
- G06N: COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06T: IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06V: IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06F: ELECTRIC DIGITAL DATA PROCESSING (computer systems based on specific computational models G06N)

## Abstract

A computerized automated parameterization image pattern detection and classification method performs (1) morphological metrics learning using labeled region data to generate morphological metrics; (2) intensity metrics learning using learning image and labeled region data to generate intensity metrics; and (3) population learning using the morphological metrics and the intensity metrics to generate learned pattern detection parameter. The method may further update the learned pattern detection parameter using additional labeled region data and learning image, and apply pattern detection with optional user parameter adjustment to image data to generate detected pattern. The method may alternatively perform pixel parameter learning and pixel classification to generate pixel class confidence, and uses the pixel class confidence and the labeled region data to perform pattern parameter learning to generate the learned pattern detection parameter. The method may further perform pattern classification learning to generate pattern classifier which is used to generate classified pattern.

## Claims

### Claim 1

A computerized automated parameterization image pattern detection method, comprising the steps of:
a) inputting learning image and labeled region data into electronic storage means;
b) performing a morphological metrics learning by computing means using the labeled region data to generate at least one morphological metric;
c) performing an intensity metrics learning by computing means using the learning image and the labeled region data to generate at least one intensity metric; and
d) performing a population learning by computing means using the at least one morphological metric and the at least one intensity metric to generate at least one learned pattern detection parameter.

### Claim 2 (depends on 1)

The computerized automated parameterization image pattern detection method of claim 1, further comprising the steps of:
a) inputting additional learning image and additional labeled region data into electronic storage means;
b) performing an updated morphological metrics learning by computing means using the at least one morphological metric and the additional labeled region data to generate at least one updated morphological metric;
c) performing an updated intensity metrics learning by computing means using the at least one intensity metric, the additional learning image and the additional labeled region data to generate at least one updated intensity metric; and
d) performing an updated population learning by computing means using the at least one learned pattern detection parameter, the at least one updated morphological metric and the at least one updated intensity metric to generate at least one updated learned pattern detection parameter.

### Claim 3 (depends on 1)

The computerized automated parameterization image pattern detection method of claim 1, further comprising the steps of:
a) inputting image data into electronic storage means; and
b) applying a pattern detection by computing means to the image data using the at least one learned pattern detection parameter to generate a detected pattern.

### Claim 4 (depends on 3)

The computerized automated parameterization image pattern detection method of claim 3, wherein the pattern detection further comprises user parameter adjustment.

### Claim 5

A computerized automated parameterization image pattern detection method, comprising the steps of:
a) inputting learning image and labeled region data into electronic storage means;
b) performing a pixel parameter learning by computing means using the labeled region data and the learning image to generate pixel classifier;
c) applying a pixel classification by computing means to the learning image using the pixel classifier to generate a pixel class confidence; and
d) performing a pattern parameter learning by computing means using the labeled region data and the pixel class confidence to generate at least one learned pattern detection parameter.

### Claim 6 (depends on 5)

The computerized automated parameterization image pattern detection method of claim 5, wherein the pixel classifier consists of a pixel feature extractor and a pixel feature classifier, and wherein the pixel parameter learning comprises the steps of:
a) performing a pixel feature extraction learning using the labeled region data and the learning image to generate the pixel feature extractor and labeled region pixel features; and
b) performing a pixel feature classification learning using the labeled region pixel features to generate the pixel feature classifier.

### Claim 7 (depends on 5)

The computerized automated parameterization image pattern detection method of claim 5, wherein the pixel classifier consists of a pixel feature extractor and a pixel feature classifier, and wherein the pixel classification comprises the steps of:
a) performing a pixel feature extraction using the learning image and the pixel feature extractor to generate pixel features; and
b) performing a pixel feature classification using the pixel features and the pixel feature classifier to generate the pixel class confidence.

### Claim 8 (depends on 5)

The computerized automated parameterization image pattern detection method of claim 5, wherein the pattern parameter learning comprises the steps of:
a) performing a morphological metrics learning by computing means using the labeled region data to generate at least one morphological metric;
b) performing an intensity metrics learning by computing means using the learning image and the pixel class confidence to generate at least one intensity metric; and
c) performing a population learning by computing means using the at least one morphological metric and the at least one intensity metric to generate the at least one learned pattern detection parameter.

### Claim 9 (depends on 5)

The computerized automated parameterization image pattern detection method of claim 5, further comprising the steps of:
a) inputting additional learning image and additional labeled region data into electronic storage means;
b) performing an updated pixel parameter learning by computing means using the pixel classifier, the additional labeled region data and the additional learning image to generate updated pixel classifier;
c) applying pixel classification by computing means to the additional learning image using the updated pixel classifier to generate updated pixel class confidence; and
d) performing an updated pattern parameter learning by computing means using the learned pattern detection parameter, the additional labeled region data and the updated pixel class confidence to generate at least one updated learned pattern detection parameter.

### Claim 10 (depends on 5)

The computerized automated parameterization image pattern detection method of claim 5, further comprising the steps of:
a) inputting image data into electronic storage means; and
b) applying pattern detection by computing means to the image data using the at least one learned pattern detection parameter to generate detected pattern.

### Claim 11 (depends on 10)

The computerized automated parameterization image pattern detection method of claim 10, wherein the pattern detection further comprises user parameter adjustment.

### Claim 12

A computerized automated parameterization image pattern detection and classification method, comprising the steps of:
a) inputting learning image and labeled region data into electronic storage means;
b) performing a pixel parameter learning by computing means using the labeled region data and the learning image to generate pixel classifier;
c) applying a pixel classification by computing means to the learning image using the pixel classifier to generate pixel class confidence;
d) performing a pattern parameter learning by computing means using the labeled region data and the pixel class confidence to generate at least one learned pattern detection parameter output;
e) applying a pattern detection by computing means to the learning image using the at least one learned pattern detection parameter to generate detected pattern; and
f) performing a pattern classification learning by computing means using the labeled region data, the learning image and the detected pattern to generate pattern classifier.

### Claim 13 (depends on 12)

The computerized automated parameterization image pattern detection and classification method of claim 12, wherein the pixel classifier consists of a pixel feature extractor and a pixel feature classifier, and wherein the pixel parameter learning comprises the steps of:
a) performing a pixel feature extraction learning using the labeled region data and the learning image to generate the pixel feature extractor and labeled region pixel features; and
b) performing a pixel feature classification learning using the labeled region pixel features to generate the pixel feature classifier.

### Claim 14 (depends on 12)

The computerized automated parameterization image pattern detection and classification method of claim 12, wherein the pixel classifier consists of a pixel feature extractor and a pixel feature classifier, and wherein the pixel classification comprises the steps of:
a) performing a pixel feature extraction using the learning image and the pixel feature extractor to generate pixel features; and
b) performing a pixel feature classification using the pixel features and the pixel feature classifier to generate the a pixel class confidence.

### Claim 15 (depends on 12)

The computerized automated parameterization image pattern detection and classification method of claim 12, wherein the pattern parameter learning comprises the steps of:
a) performing a morphological metrics learning by computing means using the labeled region data to generate at least one morphological metric;
b) performing an intensity metrics learning by computing means using the labeled region data and the pixel class confidence to generate at least one intensity metric; and
c) performing a population learning by computing means using the at least one morphological metric and the at least one intensity metric to generate the at least one learned pattern detection parameter.

### Claim 16 (depends on 12)

The computerized automated parameterization image pattern detection and classification method of claim 12, wherein the pattern detection further comprises user parameter adjustment.

### Claim 17 (depends on 12)

The computerized automated parameterization image pattern detection and classification method of claim 12, wherein the pattern classifier consists of a pattern feature extractor and a pattern feature classifier, and wherein the pattern classification learning comprises the steps of:
a) performing a pattern labeling using the labeled region data and the detected pattern to generate a labeled pattern;
b) performing a pattern feature extraction learning using the labeled pattern and the learning image to generate the pattern feature extractor and labeled pattern features; and
c) performing a pattern feature classification learning using the labeled pattern features to generate the pattern feature classifier.

### Claim 18 (depends on 12)

The computerized automated parameterization image pattern detection and classification method of claim 12, wherein the pattern classifier consists of a pattern feature extractor and a pattern feature classifier, the method further comprising the steps of:
a) inputting image data and pattern regions into electronic storage means;
b) performing a pattern feature extraction using the image data, the pattern regions and the pattern feature extractor to generate pattern features; and
c) performing a pattern feature classification using the pattern features and the pattern feature classifier to generate classified patterns.

### Claim 19 (depends on 12)

The computerized automated parameterization image pattern detection and classification method of claim 12, further comprising the steps of:
a) inputting additional learning image, additional labeled region data and additional detected pattern into electronic storage means; and
b) performing an updated pattern classification learning by computing means using the pattern classifier, the additional labeled region data, the additional learning image and the additional detected pattern to generate updated pattern classifier.