Editable semantic map with virtual camera for mobile robot learning

ChangeBridge: Patent Grants - AI & Computing (G06N)

Published March 31st, 2026

Detected April 1st, 2026

Summary

USPTO granted Patent US12592038B2 to Robert Bosch GmbH for a computer-implemented method of generating editable 3D semantic maps with virtual camera data for mobile robot learning. The patent covers foreground/background scene separation and machine learning-based background completion.

View original document View source feed page

What changed

USPTO granted Patent US12592038B2 to Robert Bosch GmbH for an 'Editable semantic map with virtual camera for mobile robot learning.' The patent covers a method for generating and modifying 3D semantic maps by separating foreground and background components, using machine learning to complete incomplete background regions, and generating virtual camera data including new image and depth data.

This is a standard patent grant notice with no compliance obligations for third parties. Companies developing autonomous robots, computer vision systems, or SLAM (Simultaneous Localization and Mapping) technologies should review this patent to assess potential licensing needs or freedom-to-operate concerns for competing products.

Source document (simplified)

← USPTO Patent Grants

Editable semantic map with virtual camera for mobile robot learning

Grant US12592038B2 Kind: B2 Mar 31, 2026

Assignee

Robert Bosch GmbH

Inventors

Cheng Zhao, Yuliang Guo, Ruoyu Wang, Xinyu Huang, Liu Ren

Abstract

A computer-implemented method and system relate to computer vision. A first semantic map of an environment is three-dimensional (3D). A foreground scene and a background scene are generated individually using the semantic data of the first semantic map. The foreground scene contains foreground components of the first semantic map. The background scene contains background components of the first semantic map. A machine learning model generates an enhanced background view by completing incomplete regions of the background components. Input data is received to modify the background components, the foreground components, or both. A second semantic map is generated in 3D using the enhanced background view, the foreground components, and the input data. The second semantic map is 3D. Virtual camera data is generated using the second semantic map. The virtual camera data includes at least new image data and corresponding new depth data.

CPC Classifications

G06T 17/05 G06T 17/20 G06T 7/194 G06T 7/20 G06T 7/70 G06T 5/50 G06T 5/60 G06T 5/77 G06T 11/60 G06T 15/04 G06T 2207/20021 G06T 2207/20081 G06T 2207/20084 G06T 2207/20016 G06T 2207/20092 G06T 2207/30244 G06T 2200/08 G06T 2200/24 G06T 2210/62 G06T 2219/2016 G06T 2219/2021 G06T 2219/2004 G06T 7/11 G06T 7/50 G06T 7/73-74 G06T 17/205 G06T 2207/30196 G06T 7/44 G06T 7/80 G06T 2215/12 G06T 2215/16 G06T 11/00 G06T 11/40 G06T 15/10 G06V 20/70 G06V 10/56 G06V 10/70 G06V 10/75 G01C 21/005 G01C 21/20 G01C 21/206 G06N 3/042 G06N 3/045 G06N 3/082 G06N 3/0895 G06N 5/00 G06N 7/001