Noise reduction method and apparatus for medical ultrasound

• US 4,771,470 A
• Filed: 11/14/1985
• Issued: 09/13/1988
• Est. Priority Date: 11/14/1985
• Status: Expired due to Fees

• Current Assignee: University of Florida Research Foundation, Incorporated (State University System of Florida)
• Original Assignee: University of Florida (State University System of Florida)
• Inventors: Oliver, Leslie H., Franklin, Bruce D., Geiser, Edward A.
• Primary Examiner(s): Boudreau, Leo H.
• Assistant Examiner(s): Daley, Donald J.
• Application Number:
  US06/797,897
• Time in Patent Office: 1,034 Days
• Field of Search: 382/22, 382/27, 382/54, 358/166, 358/282, 333/165, 333/166, 367/43, 367/45, 367/901, 128/660
• US Class Current: 382/266
• CPC Class Codes: G01S 7/52063 Sector scan display
• CPC Class Codes: G01S 7/52077 with means for elimination of unwanted signals, e.g. noise or interference
• CPC Class Codes: G06T 5/20 using local operators

Abstract

Ultrasonic images are enhanced by applying to the images asymmetric balanced gradient operators or masks which are formed to produce values which are maximized at the edges of spots having a predetermined size and location relative to the mask, which spots have a high likelihood of being specular targets. The gradient operators are further formed to produce minimum values for smaller spots corresponding to acoustic speckle. In the resultant image, the edges of specular targets are enhanced and acoustic speckle is substantially eliminated. Ultrasonic images may be processed in real time with the gradient operators in either an image processing computer or in a hardware pipeline processor.

Claims

1. A method of processing ultrasound image data to reduce acoustic speckle noise and enhance selected portions of the image, the image being formed from a plurality of pixels arranged in a spatial array and said data representing an intensity level in each pixel, the method comprising weighting the data representative of preselected ones of less than all of the pixels of an asymmetric portion of said spatial array with predetermined coefficient values, said preselected pixels and said coefficient values being selected to enhance the leading edges of spots in the image which have a size greater than a predetermined size corresponding to an average size of acoustic speckle and which have a predetermined location in the asymmetric portion;

combining the weighted data to produce a transformed intensity value;

repeating said weighting and said combining for a plurality of different asymmetric portions of the spatial array;

forming another array of pixels by using each transformed intensity value as a value of a pixel of said other array which corresponds to a central pixel of the corresponding asymmetric portion; and

displaying said other array as an image.

2. The method of claim 1, wherein said weighting comprises multiplying the intensity values of pixels in a lower central area of said asymmetric portion by positive integers and multiplying the intensity values of pixels in opposite upper corners of said asymmetric portion by negative integers, and wherein said combining comprises summing the results of said multiplications.

3. The method of claim 2, wherein said integers all have an absolute value of 1.

4. The method of claim 2, wherein the predetermined size of said spots is at least equivalent to the size of an array of 3×

2 pixels, and said predetermined location comprises said lower central area.

5. The method of claim 1, wherein said preselected pixels and said predetermined coefficients are selected such that the transformed intensity value is a function of a bi-angular gradient of intensity values of a spot in a lower central area of the asymmetric portion.

6. The method of claim 5, wherein said asymmetric portion comprises a 3×

7 array.

7. The method of claim 1, wherein said repeating comprises selecting the asymmetric portions such that successive portions overlap and such that the entire spatial array of pixels is covered.

8. The method of claim 1, wherein said image is an ultrasonic image of a portion of a human body, and said image data is in polar format.

9. A method of processing ultrasound image data to reduce acoustic speckle noise and enhance selected portions of the image, the image being formed from a plurality of pixels arranged in a spatial array and said data representing an intensity level in each pixel, the method comprising applying to the data representative of preselected pixels comprising less than all of the pixels of an asymmetric portion of the spatial array a gradient operator, the operator having a plurality of elements, each element corresponding to a different pixel of the asymmetric portion and preselected ones of the elements of the operator corresponding to said preselected pixels being assigned coefficient values, said applying comprising forming a weighted sum of the product of the coefficient values and intensity values of corresponding pixels of the asymmetric portion to produce a new intensity value representative of an intensity gradient in a predetermined area of the asymmetric portion;

the preselected pixels and the coefficient values being selected to enhance the intensity of spots in the image which are greater than a predetermined size and to minimize the intensity of other spots which are less than the predetermined size;

repeating said applying for a plurality of asymmetric portions;

forming another spatial array of pixels using the new intensity values; and

displaying said other array as an image.

10. The method of claim 9, wherein said gradient operator comprises a balanced operator formed to produce a biangular gradient value that is maximized for spots of a predetermined size located in said predetermined area of the asymmetric portion.

11. The method of claim 10, wherein said gradient operator is formed such that the new intensity values enhance the edges of said spots.

12. The method of claim 9, wherein said asymmetric portion comprises a 3×

7 array of pixels and said gradient operator comprises a 3×

7 mask, and said preselected elements comprise a first group of elements located in a lower central portion of the mask, and second and third groups of elements located in opposite upper corners of the mask.

13. The method of claim 12, wherein said first group of elements has positive coefficient values and said second and third groups have negative coefficient values.

14. The method of claim 9, wherein said repeating comprises successively applying the gradient operator to a plurality of overlapping asymmetric portions until the entire image is covered.

15. Apparatus for processing ultrasound image data to reduce acoustic speckle noise and enhance selected portions of the image, the image being formed by pluralities of pixels arranged in a plurality of lines and the image data representing the intensity level in each of the pixels, the apparatus comprising means for receiving successive data corresponding to successive pixels;

means for applying to the data an asymmetrical operator for enhancing spots in the image which correspond to specular targets and for eliminating from the image spots which correspond to acoustic speckle, the applying means comprising first delay means connected to the receiving means for delaying said data by an interval corresponding to one line;

second delay means for delaying the data from the first delay means by another interval corresponding to one line;

third delay means connected to the receiving means for delaying the data by first predetermined pixel intervals to produce a first set of delayed pixel data;

fourth delay means for delaying the data from the first delay means by second predetermined pixel intervals to produce a second set of delayed pixel data;

fifth delay means for delaying the data from the second delay means by third predetermined pixel intervals to produce a third set of delayed pixel data;

the third, fourth and fifth delay means being selected such that the first, second and third sets of delayed pixel data represent less than all of the pixels of an asymmetrical array of pixels which corresponds to said asymmetrical operator;

means for combining successive first, second, and third sets of delayed pixel data to produce corresponding successive new intensity values; and

means connected to the combining means for forming a new line of pixels from said new intensity values.

16. The apparatus of claim 15, wherein said first set of delayed pixel data comprises data corresponding to a pixel on a current line which is delayed by four pixel intervals, said second set of delayed pixel data comprises data corresponding to pixels on a previous line delayed by one, three, four, five and seven pixel intervals, and said third set of delayed pixel data comprises data corresponding to pixels on a second previous line delayed by one and seven pixel intervals.

17. The apparatus of claim 16, wherein said combining means comprises means for summing the data corresponding to the pixel on the current line and the data corresponding to the pixel on the previous line delayed by four pixel intervals to produce a result, and means for subtracting from such result the sum of the remaining data of said second and third sets to produce said new intensity value.

18. The method of claim 15, further comprising scaling said new intensity values to a predetermined range of grey level intensity values.

19. The apparatus of claim 15, further comprising means for forming a plurality of said new lines of pixels to form a processed image, and means for displaying said processed image on a video display.

20. The apparatus of claim 15, wherein said first and second delay means each comprise random access memories, and said third, fourth and fifth delay means each comprise shift registers.

21. The apparatus of claim 15, wherein said first, second and third sets of delayed pixel data are selected such that the new intensity value produced by the combining means is maximized at the edges of a spot on said image which is larger than a predetermined size.

22. The apparatus of claim 21, wherein said sets of delayed pixel data are selected such that the new intensity value corresponds to a biangular gradient of intensity values in a predetermined area of the image.