Systems and methods for block based edgel detection with false edge elimination

1. A method for generating an object edge image representation from an image data set, the method comprising:

subdividing the image data set into a plurality of image data sub-blocks, wherein the subdividing is agnostic with respect to objects represented within the image data such that the image data sub-blocks are arbitrary with respect to the objects;

separately performing edgel detection with respect to image data of each image data sub-block of the plurality of image data sub-blocks to generate a plurality of object edge image representation sub-blocks, wherein each image data sub-block of the plurality of image data sub-blocks has a corresponding object edge image representation sub-block of the plurality of object edge image representation sub-blocks; and

combining the object edge image representation sub-blocks of the plurality of object edge image representation sub-blocks using false edge elimination processing to generate the object edge image representation, wherein the combining the object edge image representation sub-blocks of the plurality of object edge image representation sub-blocks using false edge elimination processing comprises:

selecting a non-empty object edge image representation sub-block of the plurality of object edge image representation sub-blocks as a base, for the object edge image representation; and

using the false edge elimination processing to determine whether to eliminate an edgel point of a remaining object edge image representation sub-block of the plurality of object edge image representation sub-blocks or to add the edgel point of the remaining object edge image representation sub-block to the base, wherein the false edge elimination processing is iteratively applied to each remaining object edge image representation sub-block of the plurality of object edge image representation sub-blocks for combining the plurality of object edge image representation sub-blocks to generate the object edge image representation; and

wherein the false edge elimination processing comprises:

determining a surface normal value for the edgel point of the remaining object edge image representation sub-block;

determining edgel points within the base neighboring the edgel point of the remaining object edge image representation sub-block;

determining an average surface normal value for the edgel points within the base neighboring the edge point of the remaining object edge image representation sub-block; and

either removing the edgel point of the remaining object edge image representation sub-block without adding it to the base if a difference between the surface normal value and the average surface normal value is less than or equal to a threshold value or adding the edgel point of the remaining object edge image representation sub-block to the base if a difference between the surface normal value and the average surface normal value is greater than a threshold value.

2. The method of claim 1 , wherein each image data sub-block of the plurality of image data sub-blocks is of equal size.

3. The method of claim 1 , wherein the subdividing the image data set into a plurality of image data sub-blocks comprises:

determining an image volume into which the image data fits; and

setting subdivision parameters M, N, and P for controlling a number of portions each dimension of the image volume is to be divided into to provide the plurality of image data sub-blocks.

4. The method of claim 1 , wherein the separately performing edgel detection with respect to image data of each image data sub-block of the plurality of image data sub-blocks comprises:

performing edgel detection with respect to image data of two or more of the image data sub-blocks of the plurality of image data sub-blocks in parallel.

5. The method of claim 4 , wherein the performing edgel detection with respect to image data of two or more of the image data sub-blocks of the plurality of image data sub-blocks in parallel comprises:

performing edgel detection with respect to image data of each of the image data sub-blocks of the plurality of image data sub-blocks in parallel.

6. The method of claim 4 , wherein the separately performing edgel detection with respect to image data of each image data sub-block of the plurality of image data sub-blocks comprises:

performing edgel detection with respect to image data of two or more of the image data sub-blocks of the plurality of image data sub-blocks serially.

7. A system for generating an object edge image representation from an image data set, the system comprising:

at least one processor; and

memory coupled to the at least one processor, wherein the at least one processor is configured with logic to:

subdivide the image data set into a plurality of image data sub-blocks, wherein subdividing of the image data set is agnostic with respect to objects represented within the image data such that the image data sub-blocks are arbitrary with respect to the objects;

separately perform edgel detection with respect to image data of each image data sub-block of the plurality of image data sub-blocks to generate a plurality of object edge image representation sub-blocks, wherein each image data sub-block of the plurality of image data sub-blocks has a corresponding object edge image representation sub-block of the plurality of object edge image representation sub-blocks; and

combine the object edge image representation sub-blocks of the plurality of object edge image representation sub-blocks using false edge elimination processing to generate the object edge image representation, wherein the logic to combine the object edge image representation sub-blocks of the plurality of object edge image representation sub-blocks using false edge elimination processing comprises logic to:

select a non-empty object edge image representation sub-block of the plurality of object edge image representation sub-blocks as a base for the object edge image representation; and

use false edge elimination processing logic to determine whether to eliminate an edgel point of a remaining object edge image representation sub-block of the plurality of object edge image representation sub-blocks or to add the edgel point of the remaining object edge image representation sub-block to the base, wherein the false edge elimination processing is iteratively applied to each remaining object edge image representation sub-block of the plurality of object edge image representation sub-blocks for combining the plurality of object edge image representation sub-blocks to generate the object edge image representation; and

wherein the false edge elimination processing logic comprise logic to:

determine a surface normal value for the edgel point of the remaining object edge image representation sub-block:

determine edgel points within the base neighboring the edgel point of the remaining object edge image representation sub-block:

determine an average surface normal value for the edgel points within the base neighboring the edge point of the remaining object edge image representation sub-block; and

either remove the edgel point of the remaining object edge image representation, sub-block without adding it to the base if a difference between the surface normal value and the average surface normal value is less than or equal to a threshold value or add the edgel point of the remaining object edge image representation sub-block to the base if a difference between the surface normal value and the average surface normal value is greater than a threshold value.

8. The system of claim 7 , wherein each image data sub-block of the plurality of image data sub-blocks is of equal size.

9. The system of claim 7 , wherein the logic to subdivide the image data set into a plurality of image data sub-blocks comprises logic to:

determine an image volume into which the image data fits; and

set subdivision parameters M, N, and P for controlling a number of portions each dimension of the image volume is to be divided into to provide the plurality of image data sub-blocks.

10. The system of claim 7 , wherein the logic to separately perform edgel detection with respect to image data of each image data sub-block of the plurality of image data sub-blocks comprises logic to:

perform edgel detection with respect to image data of two or more of the image data sub-blocks of the plurality of image data sub-blocks in parallel.

11. The system of claim 10 , wherein the logic to perform edgel detection with respect to image data of two or more of the image data sub-blocks of the plurality of image data sub-blocks in parallel comprises logic to:

perform edgel detection with respect to image data of each of the image data sub-blocks of the plurality of image data sub-blocks in parallel.

12. The system of claim 10 , wherein the logic to separately performing edgel detection with respect to image data of each image data sub-block of the plurality of image data sub-blocks comprises logic to:

perform edgel detection with respect to image data of two or more of the image data sub-blocks of the plurality of image data sub-blocks serially.

13. A method for generating an object edge image representation from an image data set, the method comprising:

subdividing the image data set into a plurality of image data sub-blocks, wherein the subdividing is agnostic with respect to objects represented within the image data such that the image data sub-blocks are arbitrary with respect to the objects;

separately performing edgel detection with respect to image data of each image data sub-block of the plurality of image data sub-blocks to generate a plurality of object edge image representation sub-blocks, wherein each image data sub-block of the plurality of image data sub-blocks has a corresponding object edge image representation sub-block of the plurality of object edge image representation sub-blocks; and

combining the object edge image representation sub-blocks of the plurality of object edge image representation sub-blocks to generate the object edge image representation by:

selecting a non-empty object edge image representation sub-block of the plurality of object edge image representation sub-blocks as a base for the object edge image representation;

selecting a remaining non-empty object edge image representation sub-block of the plurality of object edge image representation sub-blocks for adding edgel points to the base; and

performing false edge elimination processing when combining edgel points of the selected remaining non-empty object edge image representation sub-block with the base by:

determining a surface normal value for an edgel point of the selected non-empty object edge image representation sub-block;

determining edgel points within the base neighboring the edgel point of the selected remaining non-empty object edge image representation sub-block;

determining an average surface normal value for the edgel points within the base neighboring the edge point of the selected remaining non-empty object edge image representation sub-block;

either removing the edgel point of the selected remaining non-empty object edge image representation sub-block without adding it to the base if a difference between the surface normal value and the average surface normal value is less than or equal to a threshold value or adding the edgel point of the selected remaining non-empty object edge image representation sub-block to the base if a difference between the surface normal value and the average surface normal value is greater than a threshold value; and

repeating the performing false edge elimination processing for at least each edgel point disposed on a border of the selected remaining non-empty object edge image representation sub-block;

repeating the selecting a remaining non-empty object edge image representation sub-block and performing false edge elimination processing for each remaining non-empty object edge image representation sub-block of the plurality of object edge image representation sub-blocks; and

outputting the base having edgel points from one or more of the remaining non-empty object edge image representation sub-blocks as the object edge image representation.

14. The method of claim 13 , wherein the subdividing the image data set into a plurality of image data sub-blocks comprises:

determining an image volume into which the image data fits; and

setting subdivision parameters M, N, and P for controlling a number of portions each dimension of the image volume is to be divided into to provide the plurality of image data sub-blocks.

15. The method of claim 13 , wherein the separately performing edgel detection with respect to image data of each image data sub-block of the plurality of image data sub-blocks comprises:

performing edgel detection with respect to image data of the image data sub-blocks of the plurality of image data sub-blocks in parallel.