Operator independent image cytometer
First Claim
1. An automated method of separating an object from a background in an image comprising pixels, the method comprising the steps of:
- selecting a set of weights in a digital filter, wherein at least a portion of the weights is selected to enhance at least a portion of the object, said object portion characterized by a predetermined multispectral pattern comprising a plurality of spectral bands;
transforming the image with the digital filter to enhance pixels associated with the predetermined multispectral pattern of the object not contained in the background so as to produce a transformed image, wherein the multispectral pattern comprises a plurality of edge gradients and non-edge gradients, wherein at least a portion of non-edge gradient values falls within the range of the edge gradient values, and wherein the weights are selected so as to enhance both the edge and non-edge gradients which are included in the multispectral pattern, thereby allowing object identification with the use of said digital filter; and
thresholding the transformed image, wherein the thresholding step includes extracting object features from the transformed image and sorting the extracted object features so as to separate the object from the background.
5 Assignments
0 Petitions
Accused Products
Abstract
An operator independent image cytometer having a method for image segmentation. Image segmentation comprises the steps of filtering a digital image of a cellular specimen and thresholding the resultant image. In addition, the thresholding may include the sorting of features extracted from the filtered image. The present invention also includes a method for cytometer autofocus that combines the benefits of sharpening and contrast metrics. The present invention further includes an arc lamp stabilization and intensity control system. The image cytometer has broad applications in determining DNA content and other cellular measurements on as many as 105 individual cells, including specimens of living cells. Image segmentation applications include PAP smear analysis and particle recognition.
259 Citations
29 Claims
-
1. An automated method of separating an object from a background in an image comprising pixels, the method comprising the steps of:
-
selecting a set of weights in a digital filter, wherein at least a portion of the weights is selected to enhance at least a portion of the object, said object portion characterized by a predetermined multispectral pattern comprising a plurality of spectral bands; transforming the image with the digital filter to enhance pixels associated with the predetermined multispectral pattern of the object not contained in the background so as to produce a transformed image, wherein the multispectral pattern comprises a plurality of edge gradients and non-edge gradients, wherein at least a portion of non-edge gradient values falls within the range of the edge gradient values, and wherein the weights are selected so as to enhance both the edge and non-edge gradients which are included in the multispectral pattern, thereby allowing object identification with the use of said digital filter; and thresholding the transformed image, wherein the thresholding step includes extracting object features from the transformed image and sorting the extracted object features so as to separate the object from the background. - View Dependent Claims (2, 3, 4, 5, 6, 7, 8, 9, 10, 11)
-
-
12. An operator-independent image cytometer, comprising:
-
a microscope, having a microscope stage, for optically magnifying at least a portion of a specimen having a plurality of cells; a camera for creating an image of the magnified specimen, wherein the image includes a background; means for digitizing the image so as to create a digital image; means for digital filtering the digital image into a filtered image with a filter comprising a set of weights, wherein at least a portion of the weights is selected to enhance at least a portion of the cells, said portion characterized by a multispectral pattern comprising a plurality of spectral bands so as to create contrast between the cells and the background, wherein the multispectral pattern comprises a plurality of edge gradients and non-edge gradients, wherein at least a portion of the non-edge gradient values falls within the range of edge gradient values, and wherein the weights are selected so as to enhance both the edge and non-edge gradients which are included in the multispectral pattern; and
,means for thresholding the filtered image. - View Dependent Claims (13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23)
-
-
24. An operator-independent image cytometer, comprising:
-
a microscope for creating a magnified image of a specimen; a stage controller for moving the specimen relative to the microscope; a camera for acquiring the magnified image; and means for analyzing the acquired magnified image including a pattern filter comprising a set of weights, wherein at least a portion of the weights is selected to enhance at least a portion of the image, said portion characterized by a multispectral pattern comprising a plurality of spectral bands, wherein the multispectral pattern comprises a plurality of edge gradients and non-edge gradients, wherein at least a portion of the non-edge gradient values falls within the range of edge gradient values, and wherein the weights are selected so as to enhance both the edge and non-edge gradients which are included in the multispectral pattern. - View Dependent Claims (25)
-
-
26. An image separation system, comprising:
-
means for acquiring a digital image wherein the digital image comprises a plurality of pixels indicative of a set of objects and a background; a digital filter comprising a set of weights, wherein at least a portion of the weights is selected to enhance at least a portion of at least one object, said object portion characterized by a multispectral pattern comprising a plurality of spectral bands, said filter receiving the digital image and producing a filtered image and wherein the multispectral pattern comprises a plurality of edge gradients and non-edge gradients, wherein at least a portion of the non-edge gradient values falls within the range of edge gradient values, and wherein the weights are selected so as to enhance both the edge and non-edge gradients which are included in the multispectral pattern; wherein each pixel of the digital image and the filtered image has at least one associated numeric value corresponding to a characteristic parameter of the objects and the background; wherein a difference between the numeric values associated with the pixels of the objects and the numeric values associated with the pixels of the background is greater in the filtered image than in the digital image; a histogrammer producing a histogram of the filtered image, wherein the histogram represents the filtered image numeric values; first threshold means for determining a threshold numeric value from a histogram and separating a set of thresholded objects from the filtered image; and second threshold means for determining a local threshold value for each of the thresholded objects and selectively removing a pixel from each thresholded object if the pixel has a filtered image numeric value that is on a preselected side of the local threshold value in the preselected range of values. - View Dependent Claims (27)
-
-
28. An automated method of separating an object from a background in an image comprising pixels, the method comprising the steps of:
-
selecting a set of weights in a digital filter, wherein at least a portion of the weights is selected to enhance at least a portion of the object, said object portion characterized by a predetermined multispectral pattern comprising a plurality of spectral bands; transforming the image with the digital filter to enhance pixels associated with the predetermined multispectral pattern of the object not contained in the background so as to produce a transformed image, wherein the multispectral pattern comprises a plurality of edge gradients and non-edge gradients, wherein the background comprises a plurality of background gradients, wherein at least a portion of the background gradient values falls within the range of the edge gradient values, and wherein the weights are selected so as to enhance both the edge and non-edge gradients which are included in the multispectral pattern, thereby allowing object identification with the use of said digital filter; and thresholding the transformed image, wherein the thresholding step includes extracting object features from the transformed image and sorting the extracted object features so as to separate the object from the background. - View Dependent Claims (29)
-
Specification