IP Library › Granted Patent US 12,597,152
Granted Patent B1
US 12,597,152 · App. 18/244,658 · Granted Apr 7, 2026

Parallel block matching

Inventor: Louis Marie Joseph Sugy (Munich, DE)
Assignee: NVIDIA Corporation
G06T7/593G06T2207/10012G06T2207/20021
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 12,597,152
App. No.
18/244,658
Granted
Apr 7, 2026
Kind
B1
Abstract

Apparatuses, systems, and techniques to perform block matching. In at least one embodiment, distances to objects in images are calculated based, at least in part, on overlapping portions of the images.

Claims (29)

1 . One or more processors, comprising:

circuitry to:

calculate one or more rolling sums of differences of pixel values of two or more overlapping portions of one or more objects within two or more stereoscopic images; and

cause one or more distances to the one or more objects to be calculated based, at least in part, on the one or more rolling sums.

2 . The processor of claim 1 , wherein the one or more distances to the one or more objects comprise two or more distances to the one or more objects to be calculated in parallel.

3 . The processor of claim 1 , wherein the one or more distances to the one or more objects are calculated based, at least in part, on one or more sums of differences of pixel values within the two or more overlapping portions.

4 . The processor of claim 1 , wherein the one or more distances to the one or more objects are calculated based, at least in part, on one or more sums of absolute differences of pixel values within the two or more overlapping portions.

5 . The processor of claim 1 , wherein the one or more distances to the one or more objects are calculated based, at least in part, on one or more sums of squared differences of pixel values within the two or more overlapping portions.

6 . The processor of claim 1 , wherein the one or more distances to the one or more objects are calculated based, at least in part, on one or more sums of differences of pixel values calculated based, at least in part, on one or more rolling sums.

7 . The processor of claim 1 , wherein the one or more distances to the one or more objects are calculated based, at least in part, on one or more sums of differences of pixel values calculated based, at least in part, on one or more two-dimensional rolling sums.

8 . A method, comprising:

calculating one or more rolling sums of differences of pixel values of two or more overlapping portions of one or more objects within two or more stereoscopic images; and

calculating one or more distances to the one or more objects based, at least in part, on the one or more rolling sums.

9 . The method of claim 8 , wherein the one or more distances to the one or more objects comprise two or more distances to the one or more objects calculated in parallel, based, at least in part, on dividing the two or more stereoscopic images into two or more tiles.

10 . The method of claim 8 , further comprising calculating the one or more distances to the one or more objects based, at least in part, on one or more sums of absolute differences of pixel values within the two or more overlapping portions.

11 . The method of claim 8 , further comprising calculating the one or more distances to the one or more objects based, at least in part, on one or more sums of squared differences of pixel values within the two or more overlapping portions.

12 . The method of claim 8 , further comprising dividing the two or more stereoscopic images into two or more tiles, and iteratively shifting one or more of the tiles through a predetermined range within one or more of the stereoscopic images.

13 . The method of claim 8 , further comprising calculating the one or more distances to the one or more objects based, at least in part, on one or more sums of differences of pixel values within the two or more overlapping portions calculated based, at least in part, on one or more rolling sums.

14 . The method of claim 8 , further comprising calculating the one or more distances to the one or more objects based, at least in part, on one or more sums of differences of pixel values within the two or more overlapping portions based, at least in part, on one or more two-dimensional rolling sums within one or more tiles.

15 . A system, comprising:

one or more processors to:

calculate one or more rolling sums of differences of pixel values of two or more overlapping portions of one or more objects within two or more stereoscopic images; and

cause one or more distances to the one or more objects to be calculated based, at least in part, on the one or more rolling sums; and

one or more memory devices to store the one or more rolling sums.

16 . The system of claim 15 , wherein the one or more distances to the one or more objects comprise two or more distances to the one or more objects to be calculated in parallel.

17 . The system of claim 15 , wherein the one or more distances to the one or more objects are calculated based, at least in part, on one or more sums of differences of pixel values within the two or more overlapping portions.

18 . The system of claim 15 , wherein the one or more distances to the one or more objects are calculated based, at least in part, on one or more sums of absolute differences of pixel values within the two or more overlapping portions.

19 . The system of claim 15 , wherein the one or more distances to the one or more objects are calculated based, at least in part, on one or more sums of squared differences of pixel values within the two or more overlapping portions.

20 . The system of claim 15 , wherein the one or more distances to the one or more objects are calculated based, at least in part, on one or more sums of differences of pixel values calculated based, at least in part, on one or more two-dimensional rolling sums.

Assignments (1)
ASSIGNMENT OF ASSIGNOR'S INTEREST Recorded Oct 2, 2023
From: SUGY, LOUIS MARIE JOSEPH
To: NVIDIA CORPORATION
Reel/Frame 065088/0751 →
References Cited (31)
US 6873664B1 · Clark · 2005 [cited by examiner]
US 9986623B1 · Miu · 2018 [cited by examiner]
US 10298911B2 · Shuster · 2019 [cited by examiner]
US 10354400B2 · Peleg · 2019 [cited by examiner]
US 10455219B2 · Cohen · 2019 [cited by examiner]
US 11176388B2 · Eckman · 2021 [cited by examiner]
US 11402740B2 · Eckhouse · 2022 [cited by examiner]
US 11978188B2 · Qiu · 2024 [cited by examiner]
US 12157502B2 · Wang · 2024 [cited by examiner]
US 20120071769A1 · Dunn · 2012 [cited by examiner]
US 20140152776A1 · Cohen · 2014 [cited by examiner]
US 20140153816A1 · Cohen · 2014 [cited by examiner]
US 20170188015A1 · Heidemann · 2017 [cited by examiner]
US 20180174270A1 · Burgess · 2018 [cited by examiner]
US 20190004543A1 · Kennedy · 2019 [cited by examiner]
US 20190046858A1 · Walker · 2019 [cited by examiner]
US 20200007855A1 · Cohen · 2020 [cited by examiner]
US 20200158514A1 · Moloney · 2020 [cited by examiner]
US 20200359052A1 · Tian · 2020 [cited by examiner]
US 20210019901A1 · Arai · 2021 [cited by examiner]
US 20210158081A1 · Devitt · 2021 [cited by examiner]
US 20210191251A1 · Eckhouse · 2021 [cited by examiner]
US 20220028102A1 · Chhabria · 2022 [cited by examiner]
US 20220392092A1 · Komatsu · 2022 [cited by examiner]
US 20230086983A1 · Won · 2023 [cited by examiner]
US 20230162508A1 · Xia · 2023 [cited by examiner]
US 20240364856A1 · McNamer · 2024 [cited by examiner]
US 20240394828A1 · Kwon · 2024 [cited by examiner]
US 20250203053A1 · Bormet · 2025 [cited by examiner]
WO WO2019181890A1 · 2019 [cited by examiner]
IEEE, “IEEE Standard 754-2008 (Revision of IEEE Standard 754-1985): IEEE Standard for Floating-Point Arithmetic,” Aug. 29, 2008, 70 pages. [cited by applicant]