US 12,437,515 B2
Device and method for generating training data for a machine learning system
Anna Khoreva, Stuttgart (DE); and Edgar Schoenfeld, Tuebingen (DE)
Assigned to ROBERT BOSCH GMBH, Stuttgart (DE)
Filed by Robert Bosch GmbH, Stuttgart (DE)
Filed on Sep. 13, 2022, as Appl. No. 17/943,987.
Prior Publication US 2023/0091396 A1, Mar. 23, 2023
Int. Cl. G06V 10/774 (2022.01); G06N 3/045 (2023.01); G06V 10/764 (2022.01); G06V 10/776 (2022.01); G06V 10/82 (2022.01); G06V 20/70 (2022.01)
CPC G06V 10/774 (2022.01) [G06N 3/045 (2023.01); G06V 10/764 (2022.01); G06V 10/776 (2022.01); G06V 10/82 (2022.01); G06V 20/70 (2022.01)] 13 Claims
OG exemplary drawing
 
1. A computer-implemented method for training a first machine learning system, wherein the first machine learning system is configured to generate an output characterizing a label map of an image, the method comprising the following steps:
providing a first input and a second input, wherein the first input characterizes a binary vector characterizing respective presences or absences of classes from a plurality of classes, and wherein the second input characterizes a randomly drawn value;
determining, by a first generator of the first machine learning system, an output based on the first input and the second input, wherein the output characterizes a first label map, the first label map characterizing probabilities for the classes from the plurality of classes;
determining a representation of the first label map using a global pooling operation;
training the first machine learning system based on a loss function, wherein the loss function characterizes an F1 loss, wherein the F1 loss characterizes a difference between the first input and the representation of the first label map.