Automated method and apparatus for determining characteristics of nerve fibers

• US 5,231,580 A
• Filed: 04/01/1991
• Issued: 07/27/1993
• Est. Priority Date: 04/01/1991
• Status: Expired due to Term

• Current Assignee: The United States of America As Represented By The Secretary of Agriculture
• Original Assignee: United States Department Of Health And Human Services (Government of the United States of America)
• Inventors: Cheung, Dominic, Christman, Christopher L., Thakor, Nitish V.
• Primary Examiner(s): Envall, Jr., Roy N.
• Assistant Examiner(s): Brutman, Laura
• Application Number:
  US07/678,918
• Time in Patent Office: 848 Days
• Field of Search: 364/413.13, 364/413.01, 128/774, 382/6
• US Class Current: 382/128
• CPC Class Codes: G06T 2207/30024 Cell structures in vitro; Tissue sections in vitro
• CPC Class Codes: G06T 7/60 Analysis of geometric attributes
• CPC Class Codes: G06V 20/695 Preprocessing, e.g. image segmentation

Abstract

An automated method and apparatus determines characteristics of nerve fibers. The apparatus includes an imaging unit that creates a digital image data representative of an image of a cross-section of a nerve fiber to be tested; a control unit, including image processing capability, is coupled to the imaging unit and analyzes the digital image data created by the imaging unit to identify particular characteristics of the nerve fibers including axon area, the myelin sheath area, the fiber area, the myelin sheath thickness, the fiber perimeter and the fiber eccentricity.

Claims

1. A method of determining characteristics of nerve fibers, said method comprising the steps of:

generating digital image data representative of an image of a nerve fiber cross-section using an electronic camera;

storing said digital image data in a storage unit;

accessing said digital image data from said storage unit and locating local maxima of said digital image data with a control unit;

performing a variable thresholding operation on said digital image data;

refining said local maxima;

defining a plurality of blob image regions using blob coloring;

performing an elimination operation to identify which of said plurality of blob image regions represent nerve fibers.

2. A method of determining characteristics of nerve fibers as claimed in claim 1, further comprising the step of calculating at least one of the axon area, the myelin sheath area, the fiber area, the myelin sheath thickness, the fiber perimeter and the fiber eccentricity of the blob image regions that represent nerve fibers.

3. A method of determining characteristics of nerve fibers as claimed in claim 1, further comprising the step of counting the number of blob image regions that represent nerve fibers.

4. A method of determining characteristics of nerve fibers as claimed in claim 1, wherein said local maxima are defined as a point (x,y) in an image i(x,y) such that the following conditions are met:

space="preserve" listing-type="equation">i(x,y)≧

i(x-1,y-1)

space="preserve" listing-type="equation">i(x,y)≧

i(x,y-1)

space="preserve" listing-type="equation">i(x,y)≧

i(x+1,y-1)

space="preserve" listing-type="equation">i(x,y)≧

i(x-1,y)

space="preserve" listing-type="equation">i(x,y)>

i(x+1,y)

space="preserve" listing-type="equation">i(x,y)>

i(x-1,y+1)

space="preserve" listing-type="equation">i(x,y)>

i(x,y+1)

space="preserve" listing-type="equation">i(x,y)>

i(x+1,y+1)

5. A method of determining characteristics of nerve fibers as claimed in claim 1, wherein said step of performing a variable thresholding operation includes the step of determining a plurality of local average intensities corresponding to regions of said image.

6. A method of determining characteristics of nerve fibers as claimed in claim 5, wherein said step of determining local average intensities comprises interpolation.

7. A method of determining characteristics of nerve fibers as claimed in claim 5, wherein said step of performing a variable thresholding operation comprises the further step of interpolating among said local average intensities corresponding to regions of said image.

8. A method of determining characteristics of nerve fibers as claimed in claim 1 comprising the further step of obtaining an eccentricity of said blob image regions and eliminating blob image regions as not representing axons based on eccentricity thereof.

9. A method of determining characteristics of nerve fibers, comprising the steps of:

generating digital image data representative of an image of a nerve fiber cross-section using an electronic camera;

storing said digital image data in a storage unit;

accessing said digital image data from said storage unit and locating local maxima of said digital image data with a control unit;

performing a variable thresholding operation on said digital image data;

refining said local maxima;

defining a plurality of blob image regions using blob coloring; and

performing an elimination operation to identify which of said plurality of blob image regions represent nerve fibers,

wherein said step of performing an elimination operation includes the steps of a) determining how many neighbors each of said blob image regions has and terminating the elimination operation when each of said blob image regions has only one or zero neighbors, b) deleting the blob image region with the largest number of neighbors, if there are no regions having only one neighbor, c) selecting from the blob image regions which are neighbors of blob image regions having only one neighbor and deleting the one with the most number of neighbors, and d) repeating steps a), b) and c).

10. A method of determining characteristics of nerve fibers as claimed in claim 9, wherein step a) includes the step of identifying an eccentricity of said blob image regions and deleting a blob image region having a highest eccentricity if there are two blob image regions with the same number of neighbors and step c) includes the step of deleting the blob image region with the highest eccentricity if there are two blob image regions having the same number of neighbors which are neighbors of blob image regions having only one neighbor.

11. A method of determining characteristics of nerve fibers as claimed in claim 10, further comprising the step of identifying pairs of blob image region which are sole neighbors of each other and deleting the blob image region of each pair of blob image regions that has the higher eccentricity of the pair if that blob image region has eccentricity higher than 14.

12. A method of determining characteristics of nerve fibers as claimed in claim 10, further comprising the step of identifying pairs of blob image regions which are sole neighbors of each other and keeping both blob image regions if both blob image regions have eccentricities lower than 14.

13. In a method of measuring diameters of nerve fibers using image analysis, the improvement comprising the steps of:

obtaining an eccentricity characteristic for a plurality of blob image regions and

eliminating blob image regions as not representing nerve fibers based on the eccentricity characteristics thereof.

14. The method of claim 13 wherein said eliminating step comprises the steps of determining a number of neighbors for each blob image region has, deleting a blob image region having a highest eccentricity if there are two blob image regions with the same number of neighbors and deleting the blob image region with the highest eccentricity if there are two blob image regions having the same number of neighbors which are neighbors of blob image regions having only one neighbor.

15. The method of claim 13 comprising the step of using blob coloring to define said blob image regions.