Automated cytological specimen classification system and method

• US 5,287,272 A
• Filed: 10/11/1989
• Issued: 02/15/1994
• Est. Priority Date: 04/08/1988
• Status: Expired due to Term

• Current Assignee: Tripath Imaging Incorporated (Becton, Dickinson & Co)
• Original Assignee: Neuromedical Systems, Inc.
• Inventors: Domes, Robert, Luck, Randall L., Dulak, Thomas, Porzio, John, Herriman, James M., Chaban, Richard, Knapp, James, Hall, Thomas L., Rutenberg, Mark R.
• Primary Examiner(s): Envall, Jr., Roy N.
• Assistant Examiner(s): BODENDORF, ANDREW
• Application Number:
  US07/420,105
• Time in Patent Office: 1,588 Days
• Field of Search: 364/413.1, 364/413.2, 364/413.8, 364/413.9, 364/801, 364/807, 382/6, 382/15, 382/36
• US Class Current: 382/224
• CPC Class Codes: G01N 15/1433 using image recognition
• CPC Class Codes: G01N 15/1468 with spatial resolution of the texture or inner structure of the particle
• CPC Class Codes: G01N 2015/1488 Methods for deciding
• CPC Class Codes: G06N 3/04 Architecture, e.g. interconnection topology
• CPC Class Codes: G06V 20/693 Acquisition
• CPC Class Codes: G06V 20/698 Matching; Classification
• CPC Class Codes: G16H 30/40 for processing medical images, e.g. editing
• CPC Class Codes: G16H 50/20 for computer-aided diagnosis, e.g. based on medical expert systems
• CPC Class Codes: Y10S 128/925 Neural network
• CPC Class Codes: Y10S 706/924 Medical

Abstract

An automated screening system and method for cytological specimen classification in which a neural network is utilized in performance of the classification function. Also included is an automated microscope and associated image processing circuitry.

Claims

1. A cytological specimen classifier, comprising:

a) microscope means for obtaining a view of at least part of a cytological specimen,

b) camera means for creating an image of such view,

c) image digitizing means for producing a digital representation of such image,

d) a primary classifier for identification of objects in such digital representation based on a detectable feature, and

e) secondary adaptive classifier means for recognition of cells having patterns atypical of patterns in cells expected in the specimen among the objects identified by the primary classifier.

2. The classifier of claim 1, further including tertiary classifier means for detection of atypical cells among the cells recognized by the secondary classifier.

3. The classifier of claim 2, further including a high resolution color monitor to facilitate tertiary classification of cells recognized by the secondary adaptive classifier means.

4. The classifier of claim 1, wherein said secondary adaptive classifier means includes a neural network.

5. The classifier of claim 1, further including means for reading information from a slide containing a cytological specimen.

6. The classifier of claim 5, wherein said means for reading information includes a barcode reader.

7. The classifier of claim 5, wherein said means for reading information includes an optical character reader.

8. The classifier of claim 1, further including means for marking a location of a cell recognized by the secondary adaptive classifier means on the cytogical specimen.

9. The classifier of claim 1, wherein said primary classifier includes a computer processor.

10. The classifier of claim 9, wherein said computer processor includes memory for storing digital representations of images after classification.

11. The classifier of claim 1, said secondary adaptive classifier means comprising means for recognizing cells having patterns typical of malignant or pre-malignant cells.

12. The classifier of claim 1, said secondary adaptive classifier means comprising means for recognizing cells having patterns typical of endometrial tissue.

13. The classifier of claim 1, said primary classifier comprising means for eroding and dilating objects to screen objects in such digital representation.

14. The classifier of claim 1, said primary classifier comprising means for screening objects in such digital representation on the basis of the color of the objects.

15. The classifier of claim 1, wherein said secondary adaptive classifier means includes a parallel processor.

16. The automated classifier of claim 1, said microscope means comprising an automated microscope.

17. The automated classifier of claim 1, said camera means comprising a video camera.

18. The automated classifier of claim 1, said camera means comprising a charge coupled device.

19. The automated classifier of claim 1, said primary classifier comprising a primary statistical classifier.

20. The automated classifier of claim 1, said primary classifier comprising means for detection of objects in such digital representation which have a feature that exceeds a threshold level.

21. The automated classifier of claim 1, said primary classifier comprising means for detection of objects in such digital representation which have a feature that exceeds a threshold integrated optical density.

22. The automated classifier of claim 1, said primary classifier comprising means for detection of objects in such digital representation of a cytological specimen based on morphological criteria.

23. A quality control cytological specimen classifier for classifying specimens which previously have been screened as not containing pre-malignant or malignant cells, comprising:

a) microscope means for obtaining a view of a prescreened cytological specimen,

b) camera means for creating an image of such view,

c) image digitizing means for producing a digital representation of such image,

d) a primary classifier for identification of objects in such digital representation of a cytological specimen based on a detectable feature,

e) secondary adaptive classifier means for recognition of cells having patterns typical of pre-malignant or malignant cells among the prescreened objects identified by the primary classifier, and

f) a tertiary classifier for detection of malignant and pre-malignant cells among the prescreened cells recognized by the secondary classifier.

24. A method of classifying objects in a cytological specimen, comprising the steps of:

a) obtaining a view of a cytological specimen,

b) creating an image of such view,

c) producing a digital representation of such image,

d) primarily classifying objects in such digital representation of a cytological specimen based on a detectable feature, and

e) secondarily classifying cells having features atypical of cells expected in the specimen among the objects identified in the primary classification step using adaptive processing.

25. The method of claim 24, further including the step of tertiarily classifying cells recognized by the secondary classification step for detection of unexpected cells.

26. The method of claim 24, wherein said adaptive processing includes performing classification by a neural network.

27. The method of claim 24, further including the step of marking a location of a cell recognized by the secondary classification step.