Use of machine-trained network for misalignment-insensitive depth perception
US10742959B1 · kind B1 · utility
4Cited by
8References
21Claims
0Family size
Assignee
Inventors
Key dates
| Filing date | Jan 12, 2018 |
| Grant date | Aug 11, 2020 |
| Priority date | — |
| Expiry date | Apr 17, 2038 |
Classification
- Technology area (CPC H)Electricity
- CPC primaryH04N2013/0081
- WIPO fieldAudio-visual technology
- WIPO sectorElectrical engineering
Abstract
Some embodiments of the invention provide a novel multi-layer node network to reliably determine depth based on a plurality of input sources (e.g., cameras, microphones, etc.) that may be arranged with deviations from an ideal alignment or placement. Determined depths are used, in some embodiments, to process data captured by the plurality of input sources. Other embodiments use the calculated depth to determine whether warnings must be provided or other actions taken. Some embodiments train the multi-layer network using a set of inputs generated with random misalignments incorporated into the training set.
Source: USPTO / EPO open patent data. Objective bibliographic and citation counts.