Patent · US Active

Training method for a generator neural network imposing data equivariances

US12430550B2 · kind B2 · utility

0Cited by

2References

14Claims

0Family size

Assignee

Robert Bosch GmbH · DE

Inventors

Anna Khoreva · Stuttgart, DE
Dan Zhang · San Diego, US
Edgar Schoenfeld · Amsterdam, NL

Key dates

Filing date	Jan 14, 2021
Grant date	Sep 30, 2025
Priority date	—
Expiry date	Jul 5, 2044

Classification

Technology area (CPC G)Physics
CPC primaryG06V10/82
WIPO fieldComputer technology
WIPO sectorElectrical engineering

Abstract

A training method for training a generator neural network configured to generate synthesized sensor data. A fidelity destroying transformation is defined configured to transform a measured sensor data to obtain a fidelity-destroyed transformed measured sensor data. A fidelity preserving transformation is defined configured to transform a measured sensor data to obtain a fidelity-preserved transformed measured sensor data.

Source: USPTO / EPO open patent data. Objective bibliographic and citation counts.