Method and apparatus for correcting crosstalk and spatial resolution for multichannel imaging

• US 6,763,149 B2
• Filed: 04/24/2002
• Issued: 07/13/2004
• Est. Priority Date: 04/25/2001
• Status: Expired due to Term

• Current Assignee: Cytek Biosciences Inc
• Original Assignee: Amnis Corporation (DiaSorin S.p.A.)
• Inventors: Riley, James K., Frost, Keith L.
• Primary Examiner(s): Johns, Andrew W.
• Assistant Examiner(s): Alavi, Amir
• Application Number:
  US10/132,059
• Publication Number:
  US 20030016882A1
• Time in Patent Office: 811 Days
• Field of Search: 382/203, 382/218, 382/276, 382/277, 382/282, 382/287, 382/294, 382/300, 356/73, 356/317, 356/326, 345/58
• US Class Current: 382/294
• CPC Class Codes: G01J 3/28 Investigating the spectrum using colour filters G01J3/51
• CPC Class Codes: G01J 3/2803 using photoelectric array detector
• CPC Class Codes: G01J 3/32 Investigating bands of a spectrum in sequence by a single detector
• CPC Class Codes: G01J 3/36 Investigating two or more bands of a spectrum by separate detectors
• CPC Class Codes: G01N 15/1433 using image recognition
• CPC Class Codes: G01N 15/147 the analysis being performed on a sample stream
• CPC Class Codes: G06F 2218/02 Preprocessing
• CPC Class Codes: G06T 2207/10036 Multispectral image; Hyperspectral image
• CPC Class Codes: G06T 5/50 using two or more images, e.g. averaging or subtraction
• CPC Class Codes: G06T 7/32 using correlation-based methods
• CPC Class Codes: G06T 7/33 using feature-based methods
• CPC Class Codes: G06V 20/69 Microscopic objects, e.g. biological cells or cellular parts

Abstract

A multichannel imaging system generates an ensemble of images for each field of view of an object. Each image in the ensemble is intended to contain information from only one source among a plurality of sources for the object. However, due to crosstalk, at least a portion of the signal from a first source appears in a channel intended for a second source. Because the accuracy of the correction will be degraded if the images in an ensemble are spatially misaligned with respect to one another, the spatial offset between images is determined and a correction is applied to substantially eliminate the offset. Then, a correction to the signals is determined to substantially reduce the contributions to the signal in a channel from the signals in other channels. The signal processing can be employed to process the output signals for each of a plurality of different disclosed imaging systems.

Claims

1. A method for processing signals from different channels in a multi-channel digital imaging system, said signals comprising image signal data, comprising steps of:

(a) determining spatial alignment offsets for the image signal data;

(b) aligning the image signal data by applying the spatial alignment offsets to the image signal data, to produce aligned image signal data;

(c) determining spectral crosstalk reduction coefficients; and

(d) applying the spectral crosstalk coefficients to the aligned image signal data, for spectral correction thereof.

2. The method of claim 1, wherein the step of determining spatial alignment offsets comprises the step of determining spatial alignment offsets to within sub-pixel resolution, and wherein the step of aligning the image signal data comprises the step of aligning the image signal data by applying the sub-pixel spatial alignment offsets to the image signal data.

3. The method of claim 1, wherein the step of determining spatial alignment offsets is performed before sample data are captured by the multi-channel digital imaging system.

4. The method of claim 1, wherein the step of determining spatial alignment offsets is performed concurrently with the step of capturing sample data with the multi-channel digital imaging system.

5. The method of claim 1, wherein the step of determining spatial alignment offsets comprises the steps of:

(a) selecting a first one of said different channels as a reference channel;

(b) selecting a different one of said different channels as a data channel, image signal data in the data channel to be aligned with image signal data in the reference channel;

(c) producing a correlogram by processing image signal data from the reference channel with image signal data from the data channel;

(d) determining a peak of the correlogram; and

(e) comparing the peak of correlogram with image signal data in the reference channel to determine the spatial alignment offsets.

6. The method of claim 5, wherein the step of producing a correlogram comprises the step of processing image signal data from the reference channel with image signal data from the data channel in the frequency domain.

7. The method of claim 5, wherein the step of producing a correlogram comprises the step of processing image signal data from the reference channel with image signal data from the data channel in the spatial domain.

8. The method of claim 5, wherein the step of producing a correlogram comprises the step of processing a subset of image signal data from the reference channel with a subset of image signal data from the data channel, wherein the possible spatial alignment offsets correspond to the subsets.

9. The method of claim 5, wherein the step of preparing a correlogram comprises the step of using boundary data corresponding to image signal data the reference channel and image signal data in the data channel.

10. The method of claim 9, wherein the boundary data are generated using a two dimensional gradient operator.

11. The method of claim 1, wherein the step of aligning the image signal data comprises the step of convolving the image signal data using an interpolation kernel to enable image signal data to be aligned with sub-pixel resolution.

12. The method of claim 11, wherein the interpolation kernel is determined using a function approximation of the peak of a correlogram using a Taylor series expansion.

13. The method of claim 1, wherein the step of determining spectral crosstalk reduction coefficients comprises the step of imaging a control sample, wherein said control sample comprises a source substantially limited to a single one of the different channels.

14. The method of claim 13, wherein the control sample comprises at least one of a synthesized bead and a biological sample.

15. The method of claim 1, wherein the step of determining spectral crosstalk reduction coefficients comprises the step of employing a theoretical model of a crosstalk spectrum and a sensitivity of a camera used to capture an image of a sample to a stimulus of a spectrum, in producing the image signal data.

16. The method of claim 1, wherein the step of determining spectral crosstalk reduction coefficients comprises the step of solving linear equations.

17. The method of claim 16, wherein the step of solving linear equations comprises the step of utilizing a singular value decomposition.

18. The method of claim 1, wherein the step of applying the spectral crosstalk coefficients for spectral correction comprises the step of applying a linear equation.

19. A method for correcting errors in a multichannel signal, the multichannel signal comprising an ensemble of related signals used to produce images for different channels, wherein each signal is associated with a different channel and is intended to provide information from only one source, comprising the steps of:

(a) aligning the signals in the ensemble relative to each other, such that when images produced from the signals are displayed, each image produced from a signal in the ensemble is substantially aligned with images produced from other signals in the ensemble;

(b) determining spectral crosstalk corrections suitable for correcting channel-to-channel crosstalk between signals of the ensemble; and

(c) applying the spectral crosstalk corrections to the signals associated with the different channels, to correct for the channel-to-channel crosstalk between the signals.

20. The method of claim 19, wherein the step of aligning comprises the step of applying:

(a) horizontal and vertical spatial offsets derived from a calibration signal; and

(b) constants that are accessed during the step of aligning, but which are not modified.

21. The method of claim 20, wherein the second class of constants comprises at least one of channel start columns for each signal, and inverted source coefficients.

22. The method of claim 20, wherein the step of applying the spectral crosstalk corrections comprises the step of employing constants that are accessed, but which are not modified.

23. The method of claim 19, wherein the step of aligning comprises the step of applying spatial corrections at a sub-pixel resolution.

24. The method of claim 19, further comprising the step of providing a calibration signal, wherein the step of aligning comprises the step of generating horizontal and vertical spatial offsets based upon a comparison of each signal in the ensemble with the calibration signal.

25. The method of claim 24, wherein the step of providing a calibration signal comprises the step of imaging a control sample having a single source.

26. The method of claim 24, wherein the step of providing a calibration signal further comprising the step of providing the calibration signal when the multichannel system is initialized.

27. The method of claim 24, wherein the step of providing a calibration signal further comprising the step of providing the calibration signal periodically during the use of the multichannel system.

28. The method of claim 24, wherein each signal in the ensemble comprises image data, and horizontal and vertical spatial offsets are determined for each pixel of the image data, to align the images in the different channels.

29. The method of claim 28, wherein the step of aligning, for successive signals in the ensemble that are processed, comprises the steps of:

(a) detecting a boundary of a signal currently being processed;

(b) preparing a correlogram based on the boundary and a reference signal, thereby enabling location of a peak in the correlogram;

(c) repositioning the signal currently being processed, to correspond to the peak of the correlogram.

30. The method of claim 29, wherein the step of detecting a boundary of the signal currently being processed comprises the step of using a two-dimensional gradient operator to suppress flat surfaces and to enhance object boundaries.

31. The method of claim 29, wherein the step of preparing a correlogram based on the boundary and the reference signal comprises the step of preparing a correlogram in a frequency domain.

32. The method of claim 31, wherein the step of preparing a correlogram in the frequency domain comprises the steps of:

(a) performing a Fourier Transform of boundary data for a selected signal from the ensemble and a Fourier Transform of the reference signal;

(b) multiplying a result of the Fourier Transform of the boundary data for the selected signal by a result of the Fourier Transform of the reference signal to generate a product; and

(c) performing an inverse Fourier Transform of the product.

33. The method of claim 29, wherein the step of preparing a correlogram based on the boundary and the reference signal comprises the step of preparing a correlogram in the spatial domain.

34. The method of claim 33, wherein the step of preparing a correlogram in the spatial domain comprises the steps of performing signal processing upon a subset of possible spatial alignment offsets.

35. The method of claim 29, wherein groups of image data in each channel of the multichannel system are processed together, such that a cumulative correlogram is generated for each successive channel that is processed.

36. The method of claim 29, wherein the step of aligning further comprises the step of reconstructing each signal in the ensemble by interpolating a position of the image produced with the signal to a fraction of a pixel.

37. The method of claim 36, wherein the step of reconstructing each signal comprises the step of applying a two-dimensional interpolation.

38. The method of claim 37, wherein the step of applying a two-dimensional interpolation comprises the step of computing a new amplitude value for each pixel based on a weighted sum of a group of surrounding pixels.

39. The method of claim 38, wherein the step of computing a new amplitude value for each pixel is based on a weighted sum of a group of nine pixels, eight pixels of which surround an origin pixel.

40. The method of claim 38, wherein the step of computing a new amplitude value for each pixel comprises the step of applying a matrix of coefficients to each pixel value, wherein a sum the of the coefficients in the matrix is equal to 1.0.

41. The method of claim 29, further comprising the step of determining the peak of the correlogram by employing a Taylor series expansion, eigenvalues, and eigenvectors.

42. The method of claim 24, further comprising the step of storing the signals for a period of time, wherein the step of applying spectral crosstalk corrections comprises the step of applying spectral crosstalk corrections to at least one of the signals that have been stored for the period of time.

43. The method of claim 19, wherein the step of aligning comprises the step of aligning the signal in real time.

44. The method of claim 19, further comprising the step of storing the signals for a period of time, wherein the step of aligning the signals in the ensemble comprises the step of aligning signals that have been stored for the period of time.