# US 11,009,467 B2 · Model-based methods and apparatus for classifying an interferent in specimens

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

- Number: US 11,009,467 B2
- Document ID: US11009467B2
- Status: Granted
- Kind: B2
- Application number: 15/551,562
- Filed: 2016-02-16
- Published: 2018-02-15
- Granted: 2021-05-18
- Earliest claimed: 2015-02-17
- Assignees: Siemens Healthcare Diagnostics Inc.
- Applicants: Siemens Healthcare Diagnostics Inc.
- Inventors: JinHyeong Park; Yao-Jen Chang; Wen Wu; Terrence Chen; Benjamin Pollack
- Web page: https://patentgrep.com/patent/US11009467B2

## CPC

- G01N 21/94: Investigating contamination, e.g. dust (G01N21/85 takes precedence)
- G01J 3/462: {Computing operations in or between colour spaces; Colour management systems}
- G01J 3/50: using electric radiation detectors
- G01N 21/03: Cuvette constructions
- G01N 21/25: Colour; Spectral properties, i.e. comparison of effect of material on the light at two or more different wavelengths or wavelength bands
- G01N 21/27: using photo-electric detection (G01N21/31 takes precedence){; circuits for computing concentration (logarithmic circuits G06G7/24; photometric circuits in general G01J)}
- G01N 35/00613: {Quality control}
- G06F 18/2411: based on the proximity to a decision surface, e.g. support vector machines
- G06T 7/0012: {Biomedical image inspection}
- G06V 20/695: {Preprocessing, e.g. image segmentation}
- G06V 20/698: {Matching; Classification}
- G01N 2021/5923: {Determining zones of density; quantitating spots}
- G01N 2021/5969: {Scanning of a tube, a cuvette, a volume of sample}
- G01N 2201/12: Circuits of general importance; Signal processing
- G06T 2207/10024: Color image
- G06T 2207/30024: Cell structures in vitro; Tissue sections in vitro
- G01N: INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES (measuring or testing apparatus or processes other than immunoassay, involving enzymes or microorganisms C12M, C12Q)
- G01J: MEASUREMENT OF INTENSITY, VELOCITY, SPECTRAL CONTENT, POLARISATION, PHASE OR PULSE CHARACTERISTICS OF INFRARED, VISIBLE OR ULTRAVIOLET LIGHT; COLORIMETRY; RADIATION PYROMETRY (light sources F21, H01J, H01K, H05B; investigating properties of materials by optical means G01N)
- G06F: ELECTRIC DIGITAL DATA PROCESSING (computer systems based on specific computational models G06N)
- G06T: IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06V: IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING

## Abstract

A model-based method of inspecting a specimen for presence of one or more interferent, such as Hemolysis, Icterus, and/or Lipemia (HI L) is provided. The method includes generating a pixelated image of the specimen in a first color space, determining color components (e.g., an a-value and a b-value) for pixels in the pixelated image, classifying of the pixels as being either liquid or non-liquid, defining one or more liquid regions based upon the pixels classified as liquid, and determining a presence of one or more interferent within the one or more liquid regions. The liquid classification is based upon a liquid classification model. Pixel classification may be based on a trained multiclass classifier. Interference level for the one or more interferent may be provided. Testing apparatus adapted to carry out the method are described, as are other aspects.

## Claims

### Claim 1

A method of determining a characteristic of a specimen contained within a sample container, comprising:
generating a pixelated image of the specimen, the pixelated image comprising pixels;
passing a plurality of the pixels through a liquid classification model, wherein each pixel passed through the liquid classification model is classified as liquid or non-liquid, wherein the liquid classification model is trained by graphically outlining liquid regions in multitude of examples of sample containers having different specimen conditions, occlusion by a label, and levels of serum and red blood cell portions;
defining one or more liquid regions by pixels classified as liquid; and
determining a presence of one or more interferent within the one or more liquid regions.

### Claim 2 (depends on 1)

The method of claim 1, comprising: determining an a-value and a b-value for the pixels in the pixelated image in L*a*b color space.

### Claim 3 (depends on 1)

The method of claim 1, wherein the pixelated image is subjected to gamma correction prior to the classifying.

### Claim 4 (depends on 1)

The method of claim 1, wherein the pixelated image is compiled from multiple images.

### Claim 5 (depends on 1)

The method of claim 1, wherein the liquid classification model is based upon a liquid/non-liquid detector.

### Claim 6 (depends on 1)

The method of claim 1, wherein the liquid classification model further comprises a support vector machine.

### Claim 7 (depends on 1)

The method of claim 1, wherein the determining a presence of an interferent within the one or more liquid regions is based upon a multi-class classification model generated from multiple interferent training sets.

### Claim 8 (depends on 1)

The method of claim 1, wherein the pixels in the pixelated image that have been classified as liquid are further classified by a pixel classifier as being either:
normal, or
by interferent type as being one of lipemia, hemolysis, or icterus, wherein the pixel classifier is a supervised classification model.

### Claim 9 (depends on 8)

The method of claim 8, wherein a classification as normal or by the interferent type is based upon a summation or weighting for each of the pixels classified in each class of normal, lipemia, hemolysis, or icterus.

### Claim 10 (depends on 1)

The method of claim 1, wherein the determining a presence of an interferent within the one or more liquid regions comprises:
determining an interferent type within the one or more liquid regions based upon a multi-class classification model; and
determining an interferent level of the interferent type based upon a regression model.

### Claim 11 (depends on 10)

The method of claim 10, wherein the regression model is trained for each interferent type based upon multiple training specimens exhibiting diverse interference levels.

### Claim 12 (depends on 1)

The method of claim 1, comprising analyzing the pixelated image to determine a physical dimensional characteristic of the sample container.

### Claim 13 (depends on 1)

The method of claim 1, comprising analyzing the pixelated image to determine a color of a cap on the sample container.

### Claim 14

A testing apparatus adapted to determine a characteristic of a specimen in a sample container, comprising:
one or more image capture devices configured to capture one or more images of the specimen and at least a portion of the sample container from which a pixelated image is generated;
a liquid classification model operable to classify all pixels of the pixelated image passed through the liquid classification model as being either liquid or non-liquid and define one or more liquid regions by pixels classified as liquid, wherein the liquid classification model is trained by graphically outlining liquid regions in multitude of examples of sample containers having different specimen conditions, occlusion by a label, and levels of serum and red blood cell portions;
a pixel classifier, based on a multi-class classification model, operable to determine for the pixels in the one or more liquid regions a classification as normal, hemolysis, icterus, or lipemia; and
an interferent type detector operable to detect, based on a result of the pixel classifier, one or more interferent type within the one or more liquid regions.

### Claim 15 (depends on 14)

The apparatus of claim 14, wherein the liquid classification model further comprises a support vector machine.

### Claim 16 (depends on 14)

The apparatus of claim 14, wherein the liquid classification model is generated from multiple training sets.

### Claim 17 (depends on 14)

The apparatus of claim 14, comprising an interferent level detector operable to determine an interferent level for the one or more interferent type detected.

### Claim 18 (depends on 17)

The apparatus of claim 17, wherein the one or more interferent type is at least one selected from a group consisting of:
lipemia;
hemolysis; and
icterus.

### Claim 19 (depends on 17)

The apparatus of claim 17, wherein the interferent level detector comprises a regression model that is trained for each interferent type based upon multiple training specimens exhibiting diverse interference levels.

### Claim 20

A method of determining a characteristic of a specimen contained within a sample container, comprising:
generating a pixelated image of the specimen in the sample container at a quality control station in a first color space;
converting the pixelated image from the first color space to a second color space;
determining color components for pixels in the pixelated image in the second color space;
passing a plurality of the pixels through a liquid classification model, wherein each pixel passed through the liquid classification model is classified as liquid or non-liquid, and wherein the liquid classification model is trained by graphically outlining liquid regions in multitude of examples of sample containers having different specimen conditions, occlusion by a label, and levels of serum and red blood cell portions;
defining one or more liquid regions by the pixels that are classified as liquid pixels;
classifying the liquid pixels as being one of normal, hemolysis, icterus or lipemia based upon a multi-class classification model;
determining that the one or more liquid regions is either normal or includes one or more types of interferent based upon a number of the liquid pixels in each of normal, hemolysis, icterus or lipemia; and
determining an interferent level for at least some of the one or more types of interferent based upon a regression model.

### Claim 21 (depends on 1)

The method of claim 1, wherein the liquid classification model comprises a boosting classifier.

### Claim 22 (depends on 1)

The method of claim 1, wherein the liquid classification model comprises an artificial neural network.

### Claim 23 (depends on 1)

The method of claim 1, wherein the liquid classification model comprises a tree-based classifier.

### Claim 24 (depends on 1)

The method of claim 1, wherein the liquid classification model comprises a logistic regression as a classifier.

### Claim 25 (depends on 1)

The method of claim 1, wherein the liquid is serum or plasma.

### Claim 26 (depends on 14)

The apparatus of claim 14, wherein the liquid is serum or plasma.

### Claim 27 (depends on 1)

The method of claim 1, wherein generating the pixelated image comprises:
capturing multiple images of a sample container containing a specimen, each of the multiple images captured from a different viewpoint; and
consolidating the multiple images into a pixilated image containing pixels.