US 11,861,859 B2
System and method for disparity estimation based on cost-volume attention
Mostafa El-Khamy, San Diego, CA (US); Jungwon Lee, San Diego, CA (US); and Haoyu Ren, San Diego, CA (US)
Assigned to Samsung Electronics Co., Ltd, (KR)
Filed by Samsung Electronics Co., Ltd., Gyeonggi-do (KR)
Filed on Dec. 22, 2020, as Appl. No. 17/131,129.
Claims priority of provisional application 63/092,850, filed on Oct. 16, 2020.
Claims priority of provisional application 63/083,345, filed on Sep. 25, 2020.
Claims priority of provisional application 63/065,783, filed on Aug. 14, 2020.
Prior Publication US 2022/0051426 A1, Feb. 17, 2022
Int. Cl. G06T 7/593 (2017.01); H04N 13/239 (2018.01); H04N 13/00 (2018.01)
CPC G06T 7/593 (2017.01) [G06T 2207/20021 (2013.01); G06T 2207/20081 (2013.01); H04N 13/239 (2018.05); H04N 2013/0081 (2013.01)] 19 Claims
OG exemplary drawing
 
1. A method, comprising:
extracting first feature maps from left images captured by a first camera;
extracting second feature maps from right images captured by a second camera;
calculating a matching cost based on a comparison of the first and second feature maps to generate a cost volume;
generating an attention-aware cost volume from the generated cost volume; and
aggregating the attention-aware cost volume to generate an output disparity,
wherein generating the attention-aware cost volume comprises assigning different weights for different disparity levels in the generated cost volume.