公开(公告)号：EP4343712A1

公开(公告)日：2024-03-27

申请号：EP22197404.1

申请日：2022-09-23

Applicant: Robert Bosch GmbH

Inventor： Geiger, Andreas ,  Huang, Haiwen ,  Zhang, Dan

IPC: G06V10/44 ,  G06F18/25 ,  G06V10/46 ,  G06V10/774 ,  G06V10/80 ,  G06F18/214

Abstract: Computer-implemented method for determining a dataset (T) for training and/or testing an object detector (60) comprising the steps of:
• Determining, from a first input signal ( x i ) and by means of a first model (71, 72, 73), a mid-level cue ( mlc 1 , mlc 2 , mlc 3 ), wherein the first model (71, 72, 73) is configured for determining a mid-level cue ( mlc 1 , mlc 2 , mlc 3 ) from an input signal ( x i );
• Determining, from the determined mid-level cue ( mlc 1 , mlc 2 , mlc 3 ) and by means of a second model (81, 82, 83), a first output signal ( fo ) characterizing a first object detection, wherein the second model (81, 82, 83) is configured for determining an output signal ( o 1 , o 2 , o 3 ) characterizing an object detection determined from a mid-level cue ( mlc 1 , mlc 2 , mlc 3 );
• Determining the dataset (T), wherein the dataset (T) comprises the first input signal ( x i ) and a corresponding desired output signal ( t i ), wherein the desired output signal ( t i ) characterizes the first output signal ( fo ) .

公开(公告)号：EP4418151A1

公开(公告)日：2024-08-21

申请号：EP23156748.8

申请日：2023-02-15

Applicant: Robert Bosch GmbH

Inventor： Huang, Haiwen ,  Zhang, Dan ,  Geiger, Andreas

IPC: G06F18/214 ,  G06V10/25 ,  G06V10/26 ,  G06V10/44 ,  G06V10/764 ,  G06V10/77 ,  G06V10/774 ,  G06V10/82 ,  G06V20/70 ,  G06N3/0455 ,  G06N3/0475 ,  G06N3/0895 ,  G06V20/56

CPC classification number: G06F18/2155 ,  G06V20/70 ,  G06V10/82 ,  G06V10/7753 ,  G06V10/25 ,  G06V10/26 ,  G06V10/454 ,  G06V10/764 ,  G06V10/77 ,  G06N3/0895 ,  G06N3/0475 ,  G06V20/56 ,  G06N3/045 ,  G06N3/094

Abstract: A device and a method for determining a class for at least a part of a digital image, wherein the method comprises providing (202) a classifier for a first class and a second class, determining (206) a digital image comprising an object of the second class in at least the part of the digital image, determining (208) the class for at least the part of the digital image with the classifier, wherein determining (206) the digital image comprises determining the object of the second class with a generative model depending on a label for the first class and/or depending on at least one pixel representing an object of the first class.