US 12,223,737 B2
Method for object recognition using queue-based model selection and optical flow in autonomous driving environment, recording medium and device for performing the method
Joongheon Kim, Seoul (KR); Won Joon Yun, Seoul (KR); and SooHyun Park, Incheon (KR)
Assigned to Korea University Research and Business Foundation, Seoul (KR)
Filed by Korea University Research and Business Foundation, Seoul (KR)
Filed on May 10, 2021, as Appl. No. 17/315,554.
Claims priority of application No. 10-2020-0094506 (KR), filed on Jul. 29, 2020.
Prior Publication US 2022/0036100 A1, Feb. 3, 2022
Int. Cl. G06V 20/58 (2022.01); B60W 60/00 (2020.01); G06N 3/08 (2023.01)
CPC G06V 20/58 (2022.01) [B60W 60/001 (2020.02); G06N 3/08 (2013.01); B60W 2420/403 (2013.01); B60W 2554/4049 (2020.02)] 6 Claims
OG exemplary drawing
 
1. An object recognition method using queue-based model selection and optical flow in an autonomous driving environment, the method comprising:
preprocessing data through a dense flow in a form of a matrix by calculating an optical flow of images captured consecutively over time by a sensor for an autonomous vehicle,
wherein the preprocessing of the data comprises min-max normalizing the matrix for the dense flow representing a magnitude;
generating a confidence mask by generating a vectorized confidence threshold representing a probability that there is a moving object for each cell of the preprocessed matrix;
mapping the images captured consecutively over time to the confidence mask;
determining there is a moving object in the images when a confidence of each cell mapped to the confidence mask is lower than the corresponding confidence threshold;
determining there is no object in the images when the confidence of each cell mapped to the confidence mask is higher than the corresponding confidence threshold; and
selecting an object recognition model using a tradeoff constant between object recognition accuracy and queue stability in each time unit,
wherein the selecting of the object recognition model comprises selecting an optimal object recognition model using a size of a stored queue based on Lyapunov optimization, and
wherein the selecting of the object recognition model further comprises:
selecting a fastest object recognition model as the size of the queue is larger, and
selecting a highest accuracy object recognition model as the size of the queue is smaller.