InSeGAN: a generative approach to instance segmentation in depth images
US11651497B2 · kind B2 · utility
Assignee
Inventors
Key dates
| Filing date | Mar 25, 2021 |
| Grant date | May 16, 2023 |
| Priority date | — |
| Expiry date | Jan 28, 2042 |
Classification
- Technology area (CPC G)Physics
- CPC primaryG06T2207/30128
- WIPO fieldComputer technology
- WIPO sectorElectrical engineering
Abstract
System and method for generating verisimilar images from real depth images. Train a generative adversarial neural network (GAN) by accessing test depth images having identical instances as instances of a real depth image. Input the test depth images in the generator to generate estimated depth images representing an implicit three-dimensional model of the object. Input, each estimated depth image into a discriminator to obtain a loss and into a pose encoder to obtain a matching loss. Iteratively repeat processes until the losses are minimized to a threshold, to end training. Identify the instances in the real image using the trained GAN pose encoder, to produce a pose transformation matrix for each instance in the real image. Identify pixels in the depth images corresponding to the instances of the real image and merge the pixels for the depth images to form an instance segmentation map for the real depth image.
Source: USPTO / EPO open patent data. Objective bibliographic and citation counts.