MULTISCALE EDGE DETECTION AND FIBER ENHANCEMENT USING DIFFERENCES OF ORIENTED MEANS

• US 20100202701A1
• Filed: 07/31/2008
• Published: 08/12/2010
• Est. Priority Date: 08/01/2007
• Status: Active Grant

• Current Assignee: Yeda Research and Development Co. Ltd. (Weizmann Institute of Science)
• Original Assignee: Yeda Research and Development Co. Ltd. (Weizmann Institute of Science)
• Inventors: Brandt, Achiezer, Galun, Meirav, Basri, Ronen Ezra
• Granted Patent: US 8,385,657 B2
• Time in Patent Office: Days
• US Class Current: 382/199
• CPC Class Codes: G06T 2207/10056 Microscopic image
• CPC Class Codes: G06T 2207/30024 Cell structures in vitro; Tissue sections in vitro
• CPC Class Codes: G06T 7/13 Edge detection
• CPC Class Codes: G06V 10/443 by matching or filtering
• CPC Class Codes: G06V 20/695 Preprocessing, e.g. image segmentation

Abstract

Method, apparatus and computer program product that uses a novel algorithm for edge detection suitable for both natural as well as noisy images. A scale adaptive threshold is used along with a recursive decision process to reveal the significant edges of all lengths and orientations and to localize them accurately even in low-contrast and very noisy images. Further the algorithm is use for fiber detection and enhancement by utilizing stochastic completion-like process from both sides of a fiber. The algorithm relies on an efficient multiscale algorithm for computing all “significantly different” oriented means in an image in 0(N log p), where N is the number of pixels in the image, and p is the length of the longest structure of interest. Experimental results on both natural and noisy images present confirmation of the method, apparatus and computer program product.

Claims

1. A method for edge detection in natural and noisy images comprising the steps of:

a. applying efficient multiscale elongated filters to an image to measure the difference of oriented means of various lengths and orientations as a response;

b. estimating the effect of noise on the response of such filters and obtaining a first output;

c. applying a scale adaptive threshold to the first output along with a recursive decision process to reveal the significant edges of all lengths and orientations and to localize them accurately even in low-contrast and very noisy images and obtaining a second output; and

d. saving the second output.

2. A method according to claim 1 wherein the second output is visualized.

3. A method according to claim 1 including the further step of detecting fibers in the image.

4. A method according to claim 3 wherein the detected fibers are enhanced by utilizing a stochastic completion-like process from both sides of a fiber.

5. A method according to claim 1 wherein step a includes computing all “

significantly different”

oriented means in an image in O(N log ρ

), where N is the number of pixels in the image, and p is the length of the longest structure of interest.

6. Apparatus for edge detection in natural and noisy images comprising:

a. means for applying efficient multiscale elongated filters to an image to measure the difference of oriented means of various lengths and orientations as a response;

b. means for estimating the effect of noise on the response of such filters and obtaining a first output;

c. means for applying a scale adaptive threshold to the first output along with a recursive decision process to reveal the significant edges of all lengths and orientations and to localize them accurately even in low-contrast and very noisy images and obtaining a second output; and

d. means for saving the second output.

7. Apparatus according to claim 6 including means for visualizing the second output.

8. Apparatus according to claim 6 including means for detecting fibers in the image.

9. Apparatus according to claim 8 wherein the means for detecting fibers includes means for enhancing detected fibers by a means for performing a stochastic completion-like process from both sides of a fiber.

10. Apparatus according to claim 6 including means for computing all “

significantly different”

oriented means in an image in O(N log ρ

), where N is the number of pixels in the image, and p is the length of the longest structure of interest.

11. Computer-readable medium containing program instructions for edge detection in natural and noisy images comprising program instructions for:

a. applying efficient multiscale elongated filters to an image to measure the difference of oriented means of various lengths and orientations as a response;

b. estimating the effect of noise on the response of such filters and obtaining a first output;

c. applying a scale adaptive threshold to the first output along with a recursive decision process to reveal the significant edges of all lengths and orientations and to localize them accurately even in low-contrast and very noisy images and obtaining a second output; and

d. saving the second output to memory.

12. Computer-readable medium of claim 11 containing program instructions for visualizing the second output.

13. Computer-readable medium of claim 11 containing program instructions for detecting fibers in the image.

14. Computer-readable medium of claim 13 containing program instructions for enhancing the detected fibers by a stochastic completion-like processing from both sides of a fiber.

15. Computer-readable medium of claim 11 containing program instructions for computing all “

significantly different”

oriented means in an image in O(N log ρ

), where N is the number of pixels in the image, and p is the length of the longest structure of interest.