Methods and apparatus for determining label count during specimen characterization

Patent metadata

Field	Value
Publication number	US-11313869-B2
Application number	US-201716604132-A
Country	US
Kind code	B2
Filing date	Apr 13, 2017
Priority date	Apr 13, 2017
Publication date	Apr 26, 2022
Grant date	Apr 26, 2022

Abstract

A method of characterizing a serum and plasma portion of a specimen in regions occluded by one or more labels. The characterization may be used for Hemolysis, Icterus, and/or Lipemia, or Normal detection. The method captures one or more images of a labeled specimen container including a serum or plasma portion, processes the one or more images to provide segmentation data and identification of a label-containing region, and classifying the label-containing region with a convolutional neural network (CNN) to provide a pixel-by-pixel (or patch-by-patch) characterization of the label thickness count, which may be used to adjust intensities of regions of a serum or plasma portion having label occlusion. Optionally, the CNN can characterize the label-containing region as one of multiple pre-defined label configurations. Quality check modules and specimen testing apparatus adapted to carry out the method are described, as are other aspects.

First claim

What is claimed is:

1. A characterization method, comprising:
capturing images of a specimen container including a serum or plasma portion of a specimen, the specimen container including one or more labels provided thereon;
processing the images to provide segmentation data including identification of a label-containing region;
classifying the segmentation data on the label-containing region with a convolutional neural network; and
outputting from the convolutional neural network one or more of:
per pixel data (or per patch data) on label thickness count, and
characterization of the label-containing region as one or more of pre-defined label configurations.

2. The method of claim 1 , wherein the one or more labels includes a manufacturer label, one or more barcode labels, or both.

3. The method of claim 1 , wherein the classifying the segmentation data on the label-containing region with a convolutional neural network comprises:
providing data to the convolutional neural network about the label-containing region.

4. The method of claim 1 , wherein characterization of the label-containing region as one or more of the pre-defined label configurations comprises:
determining if there is a manufacturer's label on the specimen container.

5. The method of claim 1 , wherein characterization of the label-containing region as one or more of the pre-defined label configurations comprises:
determining if there is a barcode label on the specimen container.

6. The method of claim 1 , wherein characterization of the label-containing region as one or more of the pre-defined label configurations comprises:
determining there is a manufacturer's label and a barcode label on the specimen container.

7. The method of claim 1 , wherein characterization of the label-containing region as one or more of the pre-defined label configurations comprises:
determining there is a manufacturer's label and two barcode labels on the specimen container.

8. The method of claim 1 , wherein characterization of the label-containing region as one or more of the pre-defined label configurations comprises:
determining an overall label count on the specimen container.

9. The method of claim 8 , comprising rejecting the specimen container when the label count is equal to or above a predefined threshold label count.

10. The method of claim 1 , wherein characterization of the label-containing region as one or more of the pre-defined label configurations comprises:
determining if one or more viewpoints are fully occluded by the label-containing region.

11. The method of claim 1 , wherein characterization of the label-containing region as one or more of the pre-defined label configurations comprises:
determining there is one barcode label on the specimen container, and the one barcode label covers at least some of a manufacturers label on the specimen container.

12. The method of claim 1 , wherein characterization of the label-containing region as one or more of the pre-defined label configurations comprises:
determining there are two barcode labels on the specimen container, a covered barcode label and one that is a fully-visible barcode label, wherein the fully-visible barcode label covers at least some of the covered barcode label, and the covered barcode label covers at least some of a manufacturers label on the specimen container.

13. The method of claim 1 , wherein characterization of the label-containing region as one or more of the pre-defined label configurations comprises:
determining there are three barcode labels on the specimen container, a first covered barcode label, a second covered barcode label, and one that is a fully-visible barcode label, wherein the fully-visible barcode label covers at least some of the second covered barcode label, and the second covered barcode label covers at least some of the first covered barcode label, and the covered barcode label covers at least some of a manufacturer's label on the specimen container.

14. The method of claim 1 , wherein characterization of the label-containing region as one or more of the pre-defined label configurations comprises:
determining there is one visible barcode label on the specimen container, and that other labels together with the one visible barcode label fully occlude the serum or plasma portion around an entire circumference of the specimen container.

15. The method of claim 1 , wherein the convolutional neural network includes an architecture including a convolution layer, a pooling later, and a fully-connected layer.

16. The method of claim 1 , wherein the capturing the one or more images comprises backlighting one or more viewpoints with light sources comprising one or more spectra of R, G, B, white light, IR, and near IR.

17. The method of claim 1 , wherein the capturing the one or more images is from multiple viewpoints and with multiple exposures for each of multiple spectra.

18. The method of claim 1 , wherein barcode data in the segmentation data of the label-containing region is ignored.

19. A quality check module, comprising:
a plurality of image capture devices arranged around an imaging location, and configured to capture multiple images of a specimen container, including one or more labels and containing a serum or plasma portion of a specimen, from multiple viewpoints; and
a computer coupled to the plurality of image capture devices and adapted to process image data of the multiple images, the computer configured and capable of being operated to:
capture images of the specimen container, the serum or plasma portion, and the one or more labels,
process the images to provide segmentation data including identification of a label-containing region,
classify the label-containing region with a convolutional neural network, and
output from the convolutional neural network one or more of:
per pixel data (or per patch data) on label thickness count, and
characterization of the label-containing region as one or more of pre-defined label configurations.

20. A specimen testing apparatus, comprising:
a track;
a carrier moveable on the track and configured to contain a specimen container containing a serum or plasma portion of a specimen, the specimen container including one or more labels thereon;
a plurality of image capture devices arranged around the track and configured to capture multiple images of a specimen container, the one or more labels, and the serum or plasma portion of the specimen, from multiple viewpoints; and
a computer coupled to the plurality of image capture devices and adapted to process image data of the multiple images, the computer configured and capable of being operated to:
capture images of the specimen container, the serum or plasma portion, and the one or more labels,
process the images to provide segmentation data including identification of a label-containing region,
classify the label-containing region with a convolutional neural network, and
output from the convolutional neural network one or more of:
per pixel data (or per patch data) on label thickness count, and
characterization of the label-containing region as one or more of pre-defined label configurations.

Assignees

• Siemens Healthcare Diagnostics Inc
• Tian Jiang
• Kluckner Stefan
• Sun Shanhui
• Chang Yao Jen
• Chen Terrence
• Pollack Benjamin S

Inventors

• Tian Jiang
• Kluckner Stefan
• Sun Shanhui
• Chang Yao-Jen
• Chen Terrence
• Pollack Benjamin S

Classifications

• G06V30/2247 Primary

  Characters composed of bars, e.g. CMC-7 · CPC title

• G06V30/19173

  Classification techniques · CPC title

• G06V10/82

  using neural networks · CPC title

• G06V10/454

  Integrating the filters into a hierarchical structure, e.g. convolutional neural networks [CNN] · CPC title

• G06V10/25

  Determination of region of interest [ROI] or a volume of interest [VOI] · CPC title