Methods for differential cell counts including related apparatus and software for performing same

• US 20030096324A1
• Filed: 09/11/2002
• Published: 05/22/2003
• Est. Priority Date: 09/12/2001
• Status: Abandoned Application

• Current Assignee: Vindur Technologies, Inc.
• Original Assignee: Vindur Technologies, Inc.
• Inventors: Matveev, Mikhail, Pal, Andrew Attila
• Application Number:
  US10/241,512
• Publication Number:
  US 20030096324A1
• Time in Patent Office: Days
• US Class Current: 435/7.21
• CPC Class Codes: G01N 15/1433 using image recognition
• CPC Class Codes: G01N 2015/1486 Counting the particles
• CPC Class Codes: G01N 33/5094 for blood cell populations red blood cells G01N33/80
• CPC Class Codes: G01N 35/00069 whereby the sample substrate is of the bio-disk type, i.e. having the format of an optical disk

Abstract

The present invention provides an optical method, system and software for imaging cells, in particular blood cells. In one embodiment, laboratory samples containing blood cells are deposited onto bio-discs, which are specially manufactured discs with mixing chambers that contain specific antigens to lock down various components of the blood cells. Once in the optical drive, the disc is spun and the samples and antigens are mixed with other solutions. Electromagnetic beams are then directed at the bio-disc to interact with the samples at specific capture zones and the resulting beams are collected by a detector. The information contained in the beams is then sent to a processor that produces a digital image. Various image processing methods such as binarization, background uniformization, normalization and filtering are performed to enhance cells in the investigational data for accurate counting. Other techniques are designed to correct for irregularities such as bubbles and dim cells.

Claims

1. A method of counting cells, said method comprising the steps of:

obtaining investigational data of a sample with cells;

selecting an evaluation rectangle in said investigational data;

enhancing said investigational data within said evaluation rectangle; and

counting cells within said evaluation rectangle.

2. The method of claim 1 wherein said step of selecting further comprises the step of selecting a custom size for said evaluation rectangle.

3. The method of claim 1 wherein said step of selecting selects a plurality of evaluation rectangles.

4. The method of claim 1 wherein said step of enhancing said investigational data area further comprises the steps of:

performing background illumination uniformization on said investigational data;

performing normalization on said investigational data; and

filtering said investigational data.

5. The method of claim 4 wherein said step of performing background illumination uniformization further comprises the steps of:

choosing a size for a neighborhood rectangle;

picking a point in said investigational data;

performing horizontal scanning to calculate a first sliding average for all neighbor points located within said neighborhood rectangle centered at said point;

performing vertical scanning to calculate a second sliding average for all neighbor points located within said neighborhood rectangle centered at said point;

combining said first sliding average and second sliding average to create an overall average;

reassigning the original value of said point to a resultant value calculated by obtaining the difference between said overall average and said original value and adding said difference to a background value; and

repeating said steps of performing horizontal scanning, performing vertical scanning, combining two said averages and reassigning the original value for all points in said investigational data.

6. The method of claim 4 wherein said step of performing background illumination uniformization further comprises the steps of:

performing Fourier Transform on said investigational data to produce frequency domain functions;

removing low wavelength functions from said frequency domain functions;

removing high wavelength functions from said frequency domain functions; and

performing inverse transform on said frequency domain functions to obtain a modified version of said investigational data.

7. The method of claim 4 wherein said step of performing normalization further comprises the steps of:

calculating an average and a standard deviation of the value of all points in said investigational data;

normalizing said value of all points in said investigational data using said average and said standard deviation; and

truncating said value of some points if necessary.

8. The method of claim 4 wherein said step of filtering further comprises the steps of:

choosing a size for a neighborhood rectangle;

picking a point in said investigational data;

finding all sufficiently distinct points located in said neighborhood rectangle centered at said point;

reassigning the value of said point if the number of said sufficiently distinct points is greater than a pre-determined filtering criteria; and

repeating said steps of finding all sufficiently distinct points and reassigning the value for all points in said investigational data.

9. The method of claim 4 further comprising the steps of:

removing undesirable components from said investigational data after said filtering step; and

repeating said step of performing background illumination, said step of performing normalization and said step of filtering.

13. The method of claim 1 wherein said step of counting cells shown in said evaluation rectangle further comprises the steps of:

performing convolution on said investigational data;

searching for a plurality of local maxima of said investigational data;

removing redundant local maxima from said plurality of local maxima;

declaring remaining maxima to be bright centers of cells; and

counting cells by recognizing said bright centers of cells.

14. The method of claim 13 wherein said step of performing convolution uses an indicator function that defines a circular neighborhood wherein said circular neighborhood bounds the expected size of a cell.

15. The method of claim 13 wherein said step of performing convolution uses a Gaussian indicator function.

16. The method of claim 13 wherein said step of removing redundant local maxima further comprises the steps of:

selecting a distance threshold; and

using said distance threshold to determine whether a local maxima is redundant.

17. The method of claim 13 further including a step of performing a statistical analysis comprising the steps of:

obtaining distribution of cells based of counted cells; and

estimating cell counts in areas where cells are clumped or visibility is low.

18. The method of claim 13 further comprising the steps of:

re-sampling said investigational data at a higher resolution; and

repeating said steps of performing convolution, searching for a plurality of local maxima, removing redundant local maxima, declaring remaining maxima to be bright centers of cells and counting cells by recognizing said bright centers of cells.

19. The method of claim 13 further comprising the steps of:

removing said cells counted by bright centers from said investigational data;

counting cells by recognizing dark dims; and

adding total from said step of counting cells by recognizing bright centers to total from said step of counting by recognizing dark rims.

20. The method of claim 19 wherein said step of counting cells by recognizing dark rims further comprises the steps of:

performing inversion on said investigational data;

performing a plurality of convolutions with shifted rings;

summing results from said plurality of convolutions;

finding local maxima;

declaring maxima to be centers of cells; and

counting said centers of cells.

21. The method of claim 20 wherein said step of performing a plurality of convolutions performs convolutions without shifted rings.

22. The method of claim 20 wherein said step of performing convolution uses a Gaussian indicator function.

23. The method of claim 20 wherein said step of performing convolution uses a smoothing function.

24. The method of claim 1 wherein said step of counting cells shown in said evaluation rectangle further comprises the steps of:

performing inversion on said investigational data;

performing a plurality of convolutions with shifted rings;

summing results from said plurality of convolutions;

finding local maxima;

declaring maxima to be centers of cells; and

counting said centers of cells.

25. The method of claim 24 wherein said step of performing a plurality of convolutions performs convolutions without shifted rings.

26. The method of claim 1 wherein said step of enhancing further comprises the steps of:

performing normalizing on said investigational data;

performing filtering on said investigational data;

selecting a threshold number;

performing binarization on said investigational data by determining if said investigational data differs from a set background value by a value greater than said threshold number;

performing regularization on said investigational data;

extracting one-pixel wide boundaries in said investigational data;

filling in areas defined by said one-pixel boundaries with investigational data; and

applying convolution in said filled in areas.

27. The method of claim 1 further comprising the step of displaying on a computer monitor image representation of said investigational data.

28. The method of claim 27 wherein said step of displaying further comprises the steps of:

performing fast Fourier Transform on said investigational data to generate investigational data in the frequency domain;

removing part of the spectrum in the frequency domain; and

performing inverse transform on said investigational data in the frequency domain to enhance said investigational data for display.

29. The method of claim 1 wherein said step of obtaining investigational data of a sample with cells comprises the steps of:

providing a blood sample on an optical disc surface, said surface including one or more capture zones with one or more capture agents;

loading said optical disc into an optical reader;

rotating said optical disc;

directing, from a light source, an incident beam of electromagnetic radiation to one of said capture zones;

detecting, with a detector, a resultant beam of electromagnetic radiation formed after said incident beam interact with the disc at said capture zone;

converting the detected beam into an analog output signal; and

converting said analog output signals into digital data containing cells captured at said capture zone.

30. The method of claim 29 wherein said step of converting said analog output to said digital data further comprises the steps of:

sampling amplitudes of said analog signals at fixed intervals;

recording said sampling amplitudes in an one-dimensional array;

creating a plurality of one-dimensional arrays using said steps of sampling and recording; and

combining said plurality of one-dimensional arrays to create a two-dimensional array containing digital data of said sample.

31. The method according to claim 29 wherein said optical disc is constructed with a reflective layer such that light directed to said capture is reflected to said detector.

32. The method of claim 31 where said detector is a bottom detector.

33. The method according to claim 29 wherein the optical disc is constructed such that light directed to said capture zone is transmitted through said optical disc, said disc being between said light source and said detector.

34. The method of claim 33 wherein said detector is a top detector.

35. The method of claim 33 wherein said detector is a split detector.

36. The method of claim 29 wherein said one or more capture zones are located within one or more chambers within said optical disc.

37. The method of claim 29 wherein said optical disc comprises a plurality of windows that correspond to said capture zones.

38. The method of claim 37 wherein said step of selecting evaluation rectangles step further comprises the steps of:

finding one of said plurality of windows in said investigational data; and

cropping an evaluation rectangle of standard size inside said window.

41. The method of claim 29 wherein the surface of said optical disc contains dark spots that mark the location of said captured zones.

42. The method of claim 41 wherein said step of selecting evaluation rectangles further comprises:

finding one of said dark spots in said investigational data; and

creating an evaluation rectangle of standard size with a center located at a point found by shifting a pre-determined distance from said dark spot.

43. The method of claim 42 wherein said step of finding one of said dark spots further comprises the steps of:

performing compression on said investigational data;

performing threshold evaluation on said investigational data;

performing binarization on said investigational data;

performing regularization on said investigational data;

extracting connected components from said investigational data; and

finding a component from said connected components that corresponds to a dark spot.

45. The method of claim 42 wherein said step of finding one of said dark spots further comprises the step of reading location information

49. The method of claim 1 wherein said step of obtaining investigational data of said sample with cells comprises retrieving previously stored investigational data of samples from an archive.

50. The method of claim 49 wherein said archive catalogs said stored investigational data according to characteristics of patients.

51. The method of claim 50 wherein said step of retrieving previously stored investigational data of samples further comprises the step of selecting samples matching a plurality of criteria chosen from said characteristics of patients so that a population health trends study is conducted.

52. The method of claim 1 further comprising the step of outputting results from said step of counting cells.

53. The method of claim 52 wherein said cells are white blood cells.

54. The method of claim 53 wherein said results include counts for CD4+ cells and CD8+ cells, and a ratio of CD4+ to CD8+ cells.

55. The method of claim 54 wherein said results further include counts for CD3+ cells and CD45+ cells.

56. The method of claim 1 wherein said step of counting cells further comprises the steps of:

analyzing the distribution of cells for bubble tracks;

disregarding areas with too small local cell concentration; and

recalculating cell counts.

10. The method of 9 wherein said step of removing undesirable components further comprises the steps of:

selecting a threshold value;

performing binarization on said investigational data using said threshold value;

performing regularization on said investigational data;

extracting connected components;

selecting a size threshold; and

removing components that fail to meet said size threshold.

11. The method of 10 wherein said step of performing regularization further comprises the step of performing a plurality of erosion and expansion.

12. The method of 10 wherein said step of extracting connected components further comprises the steps of:

assigning initial component numbers to all black points on said investigational data;

picking a starting point;

setting an scan direction;

scanning all points of said investigational data to reassign the component number of each of said black points to match the component number of adjacent black points;

altering said scan direction according to a set of pre-determined rules; and

repeating said steps of scanning and altering so that said component numbers of connected black points become the same.

39. The method of 37 wherein said step of finding one of said plurality of windows further comprises the steps of:

performing compression on said investigational data;

performing threshold evaluation on said investigational data;

performing binarization on said investigational data;

performing regularization on said investigational data;

extracting connected components from said investigational data; and

finding a component from said connected components that corresponds to a window.

40. The method of 39 wherein said step of extracting connected components further comprises the steps of:

assigning initial component numbers to all black points on said investigational data;

picking a starting point;

setting an scan direction;

scanning all points of said investigational data to reassign the component number of each of said black points to match the component number of adjacent black points;

altering said scan direction according to a set of pre-determined rules; and

repeating said steps of scanning and altering so that said component numbers of connected black points become the same.

44. The method of 43 wherein said step of extracting connected components further comprises the steps of:

assigning initial component numbers to all black points on said investigational data;

picking a starting point;

setting an scan direction;

scanning all points of said investigational data to reassign the component number of each of said black points to match the component number of adjacent black points;

altering said scan direction according to a set of pre-determined rules; and

repeating said steps of scanning and altering so that said component numbers of connected black points become the same.

46. The method of 29 wherein said optical disc contains computer readable location information for locating said capture zone.

47. The method of 37 further comprising the step of displaying image of said window on a computer monitor.

48. The method of 47 where said step of displaying said image of said window further comprises:

determining if said image is skewed;

finding the direction of the skew; and

correcting the skew of said image.