Method and apparatus for providing preferentially segmented digital images

• US 5,687,251 A
• Filed: 02/21/1995
• Issued: 11/11/1997
• Est. Priority Date: 02/09/1993
• Status: Expired due to Fees

• Current Assignee: Cedars-Sinai Medical Center (Cedars Sinai Health System)
• Original Assignee: Cedars-Sinai Medical Center (Cedars Sinai Health System)
• Inventors: Erler, Brian S., Marchevsky, Alberto M.
• Primary Examiner(s): Couso, Jose L.
• Assistant Examiner(s): Bella, Matthew C.
• Application Number:
  US08/391,633
• Time in Patent Office: 994 Days
• Field of Search: 382/133, 382/134, 382/257, 382/266, 382/274, 382/199
• US Class Current: 382/133
• CPC Class Codes: G06T 2207/10056 Microscopic image
• CPC Class Codes: G06T 2207/20192 Edge enhancement; Edge preservation
• CPC Class Codes: G06T 2207/30024 Cell structures in vitro; Tissue sections in vitro
• CPC Class Codes: G06T 5/40 using histogram techniques
• CPC Class Codes: G06T 5/73 Deblurring; Sharpening
• CPC Class Codes: G06T 5/92 based on global image properties
• CPC Class Codes: G06T 7/0012 Biomedical image inspection
• CPC Class Codes: G06T 7/12 Edge-based segmentation
• CPC Class Codes: G06T 7/136 involving thresholding
• CPC Class Codes: G06V 20/695 Preprocessing, e.g. image segmentation

Abstract

A method for providing a preferentially segmented digital image of a specimen image having predetermined characteristics and containing an object of interest. A specimen image is obtained through a microscope set at a magnification value and digitized to create a digital specimen image. Preferential enhancement values are determined based upon the magnification value and the specimen characteristics, and then the digital specimen image is enhanced using the preferential enhancement values to accentuate the object dark features. This creates a dark feature enhanced image which is then binarized to yield an edge mask that consists of black object silhouettes on a white background. Detection of the edges of the object contained in the edge mask is then performed to create an edge overlay that can be used to obtain morphometric measurements or combined with the digital specimen image to create a preferentially segmented digital image suitable for densitometric measurement. The resulting morphometric and densitometric data is used to accurately classify objects of interest.

Claims

1. A method for providing an edge mask from a specimen image, obtained from a specimen on a slide that is magnified at a predetermined magnification value and has predetermined characteristics, comprising the steps of:

digitizing the specimen image with an analog to digital converter to provide a digital specimen image;

selecting a first preferential enhancement value and a second preferential enhancement value based on the characteristics of the specimen image at a particular magnification value;

enhancing the digital specimen image by preferentially accentuating pixel values of objects of interest and minimizing pixel values of background objects based upon the first and second preferential enhancement values to create a dark feature enhanced image; and

binarizing the dark feature enhanced image to create an edge mask.

2. A method as defined in claim 1, further comprising the step of:

detecting the edges of the edge mask to create an edge overlay.

3. A method as defined in claim 1, further comprising the step of:

decalibrating the digital specimen image for glare and shading prior to enhancing the digital specimen image.

4. A method as defined in claim 3, wherein the step of decalibrating further comprising the steps of:

obtaining a blank bright field image from the slide;

obtaining a dark field image from the slide;

decalibrating the digital specimen image for glare and shading prior to enhancing the digital specimen image based upon the blank bright field image and the dark field image.

5. A method as defined in claim 1, further comprising the step of:

measuring the morphometric features of the edge mask.

6. A method for providing a preferentially segmented digital image representative of a specimen image with an edge overlay from a specimen stained with a dye on a slide and magnified at a predetermined magnification value, the method comprising the steps of:

interposing an optical filter between the light source and the specimen slide;

obtaining a blank bright field representation from the slide;

digitizing the blank bright field representation with an analog to digital converter to provide a blank bright field image;

obtaining a dark field representation from the slide;

digitizing the dark field representation with an analog to digital converter to provide a dark field image;

obtaining a specimen image from the slide;

digitizing the specimen image with an analog to digital converter to provide a digital specimen image;

decalibrating the digital specimen image for glare and shading based upon the blank bright field image and the dark field image to provide a decalibrated image;

determining first and second preferential enhancement values from the magnification value of the specimen image and from the characteristics of the specimen image;

enhancing the digital specimen image by preferentially accentuating pixel values of objects of interest and minimizing pixel values of background objects based upon the first and second preferential enhancement values to create a dark feature enhanced image;

binarizing the dark feature enhanced image to create an edge mask;

detecting the edges of the edge mask to create an edge overlay; and

combining the edge overlay with the decalibrated image to create a preferentially segmented digital image.

7. A method as defined in claim 6, further comprising the step of:

measuring the densitometric features of the preferentially segmented digital image.

8. An apparatus for measuring the morphometric features of an object of interest in a specimen on a slide, having predetermined characteristics, comprising:

means for converting a specimen image of the specimen magnified at a predetermined value into an electric signal;

means for digitizing the electric signal into a digital specimen image;

means for determining first and second preferential enhancement values from the magnification value of the specimen image and from the characteristics of the specimen image;

means for obtaining a blank bright field image and a dark field image from the slide;

means for decalibrating the digital specimen image for glare and shading based upon the blank bright field image and the dark field image to create a decalibrated image;

means for enhancing the decalibrated image by preferentially accentuating pixel values of objects of interest and minimizing pixel values of background objects based upon the first and second preferential enhancement values to create a dark feature enhanced image;

means for binarizing the dark feature enhanced image to create an edge mask;

means for detecting the edges of the edge mask to create an edge overlay;

means for combining the edge overlay with the decalibrated image to create a preferentially segmented digital image;

means for measuring the morphometric features of the object of interest from the edge overlay.

9. The apparatus of claim 8, further comprising:

means for analyzing the morphometric measurement to obtain a classification of the object of interest in the specimen; and

means for displaying the classification of the object of interest.

10. The apparatus of claim 8, further comprising:

means for measuring the densitometric features of the object of interest from the preferentially segmented digital image;

means for analyzing the densitometric measurements to obtain a classification of the object of interest in the specimen; and

means for displaying the classification of the object of interest.

11. A method for providing an edge mask from a specimen image, obtained from a specimen on a slide that is magnified at a predetermined magnification value and has predetermined characteristics, comprising the steps of:

digitizing the specimen image with an analog to digital converter to provide a digital specimen image;

determining a first preferential enhancement value and a second preferential enhancement value from the characteristics of the specimen image at a particular magnification value;

enhancing the digital specimen image by preferentially accentuating pixel values of objects of interest and minimizing pixel values of background objects based upon the first and second preferential enhancement values to create a dark feature enhanced image;

binarizing the dark feature enhanced image to create an edge mask;

detecting the edges of the edge mask to create an edge overlay; and

combining the edge overlay with the digital specimen image to create a preferentially segmented digital image.

12. A method as defined in claim 11, further comprising the step of:

measuring the densitometric features of the preferentially segmented digital image.

13. A method for providing an edge mask from a specimen image, obtained from a specimen on a slide that is magnified at a predetermined magnification value and has predetermined characteristics, comprising the steps of:

digitizing the specimen image with an analog to digital converter to provide a digital specimen image;

determining a first preferential enhancement value and a second preferential enhancement value from the characteristics of the specimen image at a particular magnification value;

decalibrating the digital specimen image for glare and shading prior to enhancing the digital specimen image, including the steps of

obtaining a blank bright field image from the slide;

obtaining a dark field image from the slide; and

implementing the following equation

space="preserve" listing-type="equation">New Digital Specimen Image= (Digital Specimen Image--Dark Field Value)/(Blank Bright Field Image--Dark Field Image)!*255;

enhancing the digital specimen image by preferentially accentuating pixel values of objects of interest and minimizing pixel values of background objects based upon the first and second preferential enhancement values to create a dark feature enhanced image; and

binarizing the dark feature enhanced image to create an edge mask.

14. A method for providing an edge mask from a specimen image, obtained from a specimen on a slide that is magnified at a predetermined magnification value and has predetermined characteristics, comprising the steps of:

digitizing the specimen image with an analog to digital converter to provide a digital specimen image;

determining a first preferential enhancement value and a second preferential enhancement value from the characteristics of the specimen image at a particular magnification value;

enhancing the digital specimen image by preferentially accentuating pixel values of objects of interest and minimizing pixel values of background objects based upon the first and second preferential enhancement values to create a dark feature enhanced image by implementing the following equation

space="preserve" listing-type="equation">Dark Feature Enhanced Image= (Digital Image.sup.2 /k+Digital Image).sup.n !/k.sup.n-1

where "digital image" is the digital specimen image, "k" is the first preferential enhancement value and "n" is the second preferential enhancement value; and

binarizing the dark feature enhanced image to create an edge mask.