Learning contrastive representation for semantic correspondence
US11960570B2 · kind B2 · utility
Assignee
Inventors
Key dates
| Filing date | Aug 25, 2021 |
| Grant date | Apr 16, 2024 |
| Priority date | — |
| Expiry date | Sep 12, 2042 |
Classification
- Technology area (CPC G)Physics
- CPC primaryG06V30/1444
- WIPO fieldComputer technology
- WIPO sectorElectrical engineering
Abstract
A multi-level contrastive training strategy for training a neural network relies on image pairs (no other labels) to learn semantic correspondences at the image level and region or pixel level. The neural network is trained using contrasting image pairs including different objects and corresponding image pairs including different views of the same object. Conceptually, contrastive training pulls corresponding image pairs closer and pushes contrasting image pairs apart. An image-level contrastive loss is computed from the outputs (predictions) of the neural network and used to update parameters (weights) of the neural network via backpropagation. The neural network is also trained via pixel-level contrastive learning using only image pairs. Pixel-level contrastive learning receives an image pair, where each image includes an object in a particular category.
Source: USPTO / EPO open patent data. Objective bibliographic and citation counts.