EFFICIENT SIMD IMPLEMENTATION OF 3X3 NON MAXIMA SUPPRESSION OF SPARSE 2D IMAGE FEATURE POINTS

• US 20200134356A1
• Filed: 12/30/2019
• Published: 04/30/2020
• Est. Priority Date: 04/19/2016
• Status: Active Grant

• Current Assignee: Texas Instruments, Inc.
• Original Assignee: Texas Instruments, Inc.
• Inventors: Poddar, Deepak Kumar, Swami, Pramod Kumar, Viswanath, Prashanth
• Granted Patent: US 11,010,631 B2
• Time in Patent Office: Days
• CPC Class Codes: G06V 10/462 Salient features, e.g. scale invariant feature transforms [SIFT]
• CPC Class Codes: G06V 10/955 using specific electronic processors

Abstract

In accordance with disclosed embodiments, an image processing method includes performing a first scan in a first direction on a first list of pixels in which, for each pixel in the first list, a feature point property is compared with a corresponding feature point property of each of a first set of neighboring pixels, performing a second scan in a second direction on the first list of pixels in which, for each pixel in the first list, a feature point property is compared with a corresponding feature point property of each of a second set of neighboring pixels, using the results of the first and second scans to identify pixels from the first list to be suppressed, and forming a second list of pixels that includes pixels from the first list that are not identified as pixels to be suppressed. The second list represents a non-maxima suppressed list.

Claims

1. A method comprising:

scanning, in a first direction, a list of feature points of an image, wherein each feature point corresponds to a pixel of the image and the scanning in the first direction includes comparing, for each feature point of the list, a property of the feature point to a corresponding property of each of a first set of feature points;

scanning, in a second direction, the list of feature points of the image, wherein the scanning in the second direction includes comparing, for each feature point of the list, a property of the feature point to a corresponding property of each of a second set of feature points;

in response to the scanning in the first direction and the scanning in the second direction, identifying which feature points of the list are to be suppressed; and

producing a non-maxima suppressed feature point list that includes the feature points from the list other than the feature to be suppressed.

2. The method of claim 1, wherein the property is a combined y-coordinate and reliability score.

3. The method of claim 1, wherein the first direction is a forward direction and the second direction is a backward direction.

4. The method of claim 1, wherein the pixels corresponding to the first set of feature points differ from the pixels corresponding to the second set of feature points.

5. The method of claim 1, wherein the pixel corresponding to the feature point, the pixels corresponding to the first set of feature points, and the pixels corresponding to the second set of feature points collectively are a 3 by 3 block of pixels.

6. The method of claim 5, wherein the first set of pixels includes:

a first pixel that is a left neighbor pixel to the pixel to which the feature point corresponds;

a second pixel that is a top left neighbor pixel to the pixel to which the feature point corresponds;

a third pixel that is a top neighbor pixel to the pixel to which the feature point corresponds; and

a fourth pixel that is a top right neighbor pixel to the pixel to which the feature point corresponds; and

wherein the second set of pixels includes;

a fifth pixel that is a right neighbor pixel to the pixel to which the feature point corresponds;

a sixth pixel that is a bottom right neighbor pixel to the pixel to which the feature point corresponds;

a seventh pixel that is a bottom neighbor pixel to the pixel to which the feature point corresponds; and

an eighth pixel that is a bottom left neighbor pixel to the pixel to which the feature point corresponds.

7. A method comprising:

performing a first scan in a first direction on a first list of pixels in which, for each pixel in the first list, a feature point property is compared with a corresponding feature point property of each of a first set of neighboring pixels;

performing a second scan in a second direction on the first list of pixels in which, for each pixel in the first list, a feature point property is compared with a corresponding feature point property of each of a second set of neighboring pixels;

using the results of the first and second scans to identify pixels from the first list to be suppressed; and

forming a second list of pixels that includes pixels from the first list that are not identified as pixels to be suppressed.

8. The method of claim 7, wherein the first direction is a forward direction and the second direction is a backward direction.

9. The method of claim 7, wherein none of the pixels of the first set of neighboring pixels are part of the second set of neighboring pixels.

10. The method of claim 7, wherein the pixel, the first set of neighboring pixels, and the second set of neighboring pixels form a 3 by 3 block of pixels.

11. The method of claim 10, wherein the first set of neighboring pixels includes:

a first pixel that is a left neighbor pixel to the pixel having the feature point property being compared;

a second pixel that is a top left neighbor pixel to the pixel having the feature point property being compared;

a third pixel that is a top neighbor pixel to the pixel having the feature point property being compared; and

a fourth pixel that is a top right neighbor pixel to the pixel having the feature point property being compared; and

wherein the second set of neighboring pixels includes;

a fifth pixel that is a right neighbor pixel to the pixel having the feature point property being compared;

a sixth pixel that is a bottom right neighbor pixel to the pixel having the feature point property being compared;

a seventh pixel that is a bottom neighbor pixel to the pixel having the feature point property being compared; and

an eighth pixel that is a bottom left neighbor pixel to the pixel having the feature point property being compared.

12. An electronic device comprising:

a processor;

a memory to store image data and program instructions that, when executed by the processor, cause the processor to;

perform a first scan in a first direction on a first list of pixels in which, for each pixel in the first list, a feature point property is compared with a corresponding feature point property of each of a first set of neighboring pixels;

perform a second scan in a second direction on the first list of pixels in which, for each pixel in the first list, a feature point property is compared with a corresponding feature point property of each of a second set of neighboring pixels;

use the results of the first and second scans to identify pixels from the first list to be suppressed; and

form a second list of pixels that includes the pixels from the first list that are not identified as pixels to be suppressed.

13. The electronic device of claim 12, wherein the feature point property is a combined y-coordinate and reliability score.

14. The electronic device of claim 12, wherein the first direction is a forward direction and the second direction is a backward direction.

15. The electronic device of claim 12, wherein none of the pixels of the first set of neighboring pixels are part of the second set of neighboring pixels.

16. The electronic device of claim 12, wherein the pixel, the first set of neighboring pixels, and the second set of neighboring pixels form a 3 by 3 block of pixels.

17. The electronic device of claim 16, wherein the first set of neighboring pixels includes:

a first pixel that is a left neighbor pixel to the pixel having the feature point property being compared;

a second pixel that is a top left neighbor pixel to the pixel having the feature point property being compared;

a third pixel that is a top neighbor pixel to the pixel having the feature point property being compared; and

a fourth pixel that is a top right neighbor pixel to the pixel having the feature point property being compared; and

wherein the second set of neighboring pixels includes;

a fifth pixel that is a right neighbor pixel to the pixel having the feature point property being compared;

a sixth pixel that is a bottom right neighbor pixel to the pixel having the feature point property being compared;

a seventh pixel that is a bottom neighbor pixel to the pixel having the feature point property being compared; and

an eighth pixel that is a bottom left neighbor pixel to the pixel having the feature point property being compared.