US 12,430,728 B2
Method for labeling image object and circuit system
Yen-Ying Chen, Hsinchu (TW); Wei-Yuan Hsu, Hsinchu (TW); and Chia-Wei Yu, Hsinchu (TW)
Assigned to REALTEK SEMICONDUCTOR CORP., Hsinchu (TW)
Filed by REALTEK SEMICONDUCTOR CORP., Hsinchu (TW)
Filed on Apr. 14, 2023, as Appl. No. 18/134,625.
Claims priority of application No. 111114577 (TW), filed on Apr. 18, 2022.
Prior Publication US 2023/0334629 A1, Oct. 19, 2023
Int. Cl. G06K 9/00 (2022.01); G06T 5/77 (2024.01); G06T 7/11 (2017.01); G06T 7/55 (2017.01); G06V 10/74 (2022.01); G06V 10/764 (2022.01); G06V 10/82 (2022.01); G06V 20/70 (2022.01)
CPC G06T 5/77 (2024.01) [G06T 7/11 (2017.01); G06T 7/55 (2017.01); G06V 10/761 (2022.01); G06V 10/764 (2022.01); G06V 10/82 (2022.01); G06V 20/70 (2022.01); G06T 2207/20084 (2013.01)] 20 Claims
OG exemplary drawing
 
1. A method for labeling an image object, which is performed in a circuit system, comprising:
receiving an image;
using an object classification method to segment the image into one or more regions, classifying each of the regions into a classification, and assigning a classification label to each of the regions;
using a depth estimation method to estimate a depth of each of pixels of the image;
determining, according to the classification of each of the regions, whether or not the depth of each of the pixels in the region matches the classification of the region to which the pixel belongs; and
assigning, by the circuit system, weights to the regions according to the classification label of each of the regions when the depth of any one of the pixels in the region matches the classification of the region to which the pixel belongs, and performing a post-processing process upon the image based on the weights assigned to the regions;
wherein any of the pixels is regarded as noise when the depth of the pixel does not match the classification of the region to which the pixel belongs.