DIGITAL IMAGE EDGE DETECTION

• US 20150206319A1
• Filed: 07/10/2014
• Published: 07/23/2015
• Est. Priority Date: 01/17/2014
• Status: Active Grant

• Current Assignee: Microsoft Technology Licensing LLC (Microsoft Corporation)
• Original Assignee: Microsoft Corporation
• Inventors: Dollar, Piotr, Zitnick, Charles Lawrence III
• Granted Patent: US 9,934,577 B2
• Time in Patent Office: Days
• US Class Current: 1/1
• CPC Class Codes: G06F 18/29 Graphical models, e.g. Bayesian networks
• CPC Class Codes: G06T 2207/20081 Training; Learning
• CPC Class Codes: G06T 7/13 Edge detection
• CPC Class Codes: G06V 10/44 Local feature extraction by analysis of parts of the pattern, e.g. by detecting edges, contours, loops, corners, strokes or intersections; Connectivity analysis, e.g. of connected components
• CPC Class Codes: G06V 10/84 using probabilistic graphical models from image or video features, e.g. Markov models or Bayesian networks

Abstract

Edges are detected in a digital image including a plurality of pixels. For each of the plurality of pixels, a plurality of different edge assessments are made for that pixel. Each different edge assessment considers that pixel in a different position of a different pixel patch. The different edge assessments for each pixel are aggregated.

Claims

1. A method of detecting edges, comprising:

receiving an image including a plurality of pixels;

for each of the plurality of pixels, making a plurality of different edge assessments for that pixel, each different edge assessment considering that pixel in a different position of a different pixel patch; and

for each of the plurality of pixels, aggregating the different edge assessments for that pixel.

2. The method of claim 1, wherein each different edge assessment uses a decision tree on a pixel patch.

3. The method of claim 2, wherein a set of decision trees are used on the different pixel patches in a repeating pattern.

4. The method of claim 2, wherein the decision tree is previously trained via supervised machine learning.

5. The method of claim 1, wherein each different edge assessment uses a structured forest on a pixel patch, the structured forest using one or more decision trees on the pixel patch.

6. The method of claim 5, wherein each different decision tree is previously trained via supervised machine learning.

7. The method of claim 5, wherein the structured forest analysis uses four or fewer decision tress.

8. The method of claim 5, wherein the structured forest analysis uses between two and four different decision trees.

9. The method of claim 5, wherein different structured forests are used on different pixel patches.

10. The method of claim 9, wherein a same structured forest is periodically used for different pixel patches.

11. The method of claim 5, wherein a set of the different structured forests are used on the different pixel patches in a repeating pattern.

12. A method of detecting edges, comprising:

receiving an image including a plurality of pixels;

for each of the plurality of pixels, making a plurality of different edge assessments for that pixel, each different edge assessment considering that pixel in a different position of a different pixel patch, and each different edge assessment utilizing a structured forest including two or more decision trees, wherein different structured forests are used on different pixel patches; and

for each of the plurality of pixels, aggregating the different edge assessments for that pixel.

13. The method of claim 12, wherein a same structured forest is periodically used for different pixel patches.

14. The method of claim 12, wherein a set of the different structured forests are used on the different pixel patches in a repeating pattern.

15. The method of claim 12, wherein each decision tree is previously trained via supervised machine learning.

16. A method of detecting edges, comprising:

receiving an image including a plurality of pixels;

for each of the plurality of pixels, making a plurality of different edge assessments for that pixel, each different edge assessment considering that pixel in a different position of a different pixel patch, and each different edge assessment utilizing a structured forest including two to four decision trees; and

for each of the plurality of pixels, aggregating the different edge assessments for that pixel.

17. The method of claim 16, wherein each different decision tree is previously trained via supervised machine learning.

18. The method of claim 16, wherein different structured forests are used on different pixel patches.

19. The method of claim 18, wherein a same structured forest is periodically used for different pixel patches.

20. The method of claim 18, wherein a set of the different structured forests are used on the different pixel patches in a repeating pattern.