CPC G06V 20/56 (2022.01) [B60W 50/00 (2013.01); G06F 18/2148 (2023.01); G06F 18/241 (2023.01); G06V 10/22 (2022.01); G06V 10/25 (2022.01); G06V 10/7715 (2022.01); G06V 20/647 (2022.01); H04N 13/111 (2018.05)] | 20 Claims |
1. An advanced driver assist system (ADAS), comprising:
a processing circuit; and
a memory configured to store instructions executable by the processing circuit,
wherein the processing circuit is configured to execute the instructions to cause the ADAS to perform operations comprising:
generating an event detection model by training a trained model including a first trained model and a second trained model based on a stereo image including at least one object;
generating a position information associated with the at least one object based on reflected signals by using the first trained model;
determining regions, each comprising at least a portion of the at least one object, as candidate bounding boxes based on the stereo image by using the first trained model; and
selectively adjusting class scores of the candidate bounding boxes associated with the at least one object based on the position information by using the first trained model.
|
16. An advanced driver assist system (ADAS) in a vehicle, the ADAS comprising:
a pre-processor configured to generate a stereo image including at least one object, based on a video sequence comprising a plurality of frames;
a tracking list generation engine configured to generate a position information associated with the at least one object based on reflected signals; and
an object detector configured to train a trained model including a first trained model and a second trained model based on an input image, determine regions, each comprising at least a portion of the at least one object, as candidate bounding boxes based on the stereo image by using the first trained model, and selectively adjust class scores of the candidate bounding boxes associated with the at least one object based on the position information by using the first trained model.
|
19. A method of detecting an object in an advanced driver assist system (ADAS), the method comprising:
generating an event detection model by training a trained model including a first trained model and a second trained model based on a stereo image including at least one object,
generating a position information associated with the at least one object based on reflected signals by using the first trained model;
determining regions, each comprising at least a portion of the at least one object, as candidate bounding boxes based on the stereo image by using the first trained model; and
selectively adjusting class scores of the candidate bounding boxes associated with the at least one object based on the position information by using the first trained model.
|