IP Library › Granted Patent US 9,799,124
Granted Patent B2
US 9,799,124 · App. 14/470,271 · Granted Oct 24, 2017

Method, system and apparatus for determining a property of an image

Inventors: Geoffrey Richard Taylor (Carlingford, AU); Fei Mai (Marsfield, AU)
Assignee: CANON KABUSHIKI KAISHA
G06T7/536G06K9/00771G06K9/00208G06K9/4638G06T2207/20072
View Patent ↗
Loading inventors, assignments & file history…
Monitor This Case
Get email alerts when status or documents change.
Order Certified Copies
Most orders are placed with the USPTO same day — all within 24 business hours.
Order via The Patent Place →
Pre-filled with this patent's details
Quick Facts
Patent No.
US 9,799,124
App. No.
14/470,271
Granted
Oct 24, 2017
Kind
B2
Abstract

A method of determining a property of an image ( 176 ) captured by a camera ( 127 ). Vanishing points ( 320, 330, 340 ) are in the image ( 176 ). Each pixel of the image ( 176 ) is associated with one or more of the vanishing points ( 320, 330, 340 ) based on an orientation of the image gradient at the pixel. The image is partitioned into a set of regions associated with a pair of the determined vanishing points based on the vanishing point associations for pixels in the image ( 176 ). Boundaries of the regions are aligned with the associated vanishing points. For at least one of the plurality of regions, a confidence value is determined for the property of the image based on pixels in the region. The property of the image is determined for one or more pixels within the image ( 176 ) based on the confidence value.

Claims (38)

1. A method of determining a property of an image captured by a camera, the method comprising:

determining a plurality of vanishing points from the image;

associating each pixel of the image with at least one of the vanishing points based on alignment of a pixel edge orientation with a vanishing point direction for each pixel, a similarity of an orientation of an image gradient at each pixel, and a vanishing point direction from each pixel to each of the vanishing points;

partitioning the image into a set of regions based on a pair of the vanishing points;

constructing a space partition tree by iterating the partitioning;

determining a boundary line for each region of the set of regions that passes through at least one of the vanishing points of said pair of vanishing points;

determining, for at least one of said regions, a confidence value for the property of the image based on pixels in the at least one region; and

determining the property of the image for one or more pixels within said image based on the confidence value.

2. The method of claim 1 , wherein said property is surface orientation supported by the pair of vanishing points.

3. The method of claim 1 , wherein said property is a semantic label based on an appearance of pixels within one of the set of regions.

4. The method of claim 1 , wherein the set of regions is reduced to a second set of regions by merging neighbouring regions according to a second property.

5. A system for determining a property of an image captured by a camera, the system comprising:

a memory for storing data and a computer program; and

a processor coupled to the memory for executing the computer program, the computer program comprising instructions for:

determining a plurality of vanishing points from the image;

associating each pixel of the image with at least one of the vanishing points based on alignment of a pixel edge orientation with a vanishing point direction for each pixel, a similarity of an orientation of an image gradient at each pixel, and a vanishing point direction from each pixel to each of the vanishing points;

partitioning the image into a set of regions based on a pair of the vanishing points;

constructing a space partition tree by iterating the partitioning;

determining a boundary line for each region of the set of regions that passes through at least one of the vanishing points of said pair of vanishing points;

determining, for at least one of said regions, a confidence value for the property of the image based on pixels in the at least one region; and

determining the property of the image for one or more pixels within said image based on the confidence value.

6. A non-transitory computer readable medium having a computer program stored thereon for determining a property of an image captured by a camera, the program comprising:

code for determining a plurality of vanishing points from the image;

code for associating each pixel of the image with at least one of the vanishing points based on alignment of a pixel edge orientation with a vanishing point direction for each pixel, a similarity of an orientation of an image gradient at each pixel, and a vanishing point direction from each pixel to each of the vanishing points;

code for partitioning the image into a set of regions based on a pair of the vanishing points;

code for constructing a space partition tree by iterating the partitioning;

code for determining a boundary line for each region of the set of regions that passes through at least one of the vanishing points of said pair of vanishing points;

code for determining, for at least one of said regions, a confidence value for the property of the image based on pixels in the at least one region; and

code for determining the property of the image for one or more pixels within said image based on the confidence value.

7. An apparatus for determining a property of an image captured by a camera, the apparatus comprising:

module for determining a plurality of vanishing points from the image;

module for associating each pixel of the image with at least one of the vanishing points based on alignment of a pixel edge orientation with a vanishing point direction for each pixel, a similarity of an orientation of an image gradient at each pixel, and a vanishing point direction from each pixel to each of the vanishing points;

module for partitioning the image into a set of regions based on a pair of the vanishing points;

module for constructing a space partition tree by iterating the partitioning;

module for determining a boundary line for each region of the set of regions that passes through at least one of the vanishing points of said pair of vanishing points;

module for determining, for at least one of said regions, a confidence value for the property of the image based on pixels in the at least one region; and

module for determining the property of the image for one or more pixels within said image based on the confidence value; and

at least one processor, wherein each module is implemented by the at least one processor.

Assignments (1)
ASSIGNMENT OF ASSIGNOR'S INTEREST Recorded Aug 27, 2014
From: TAYLOR, GEOFFREY RICHARD; MAI, FEI
To: CANON KABUSHIKI KAISHA
Reel/Frame 033645/0493 →
Priority Claims (1)
AU 2013222016 · Aug 30, 2013 · national
Continuity (1)
Related Publication 20150063684A1 · Mar 5, 2015