Detecting organization image-borne sensitive documents and protecting against loss of the sensitive documents
US10867073B1 · kind B1 · utility
Assignee
Inventors
Key dates
| Filing date | Jun 3, 2020 |
| Grant date | Dec 15, 2020 |
| Priority date | — |
| Expiry date | Jun 3, 2040 |
Classification
- Technology area (CPC G)Physics
- CPC primaryG06N20/00
- WIPO fieldComputer technology
- WIPO sectorElectrical engineering
Abstract
Disclosed is customizing a DL stack to detect organization sensitive data in images and protecting against loss of the sensitive documents, pre-training a master DL stack by forward inference and back propagation using labelled ground truth data for image-borne sensitive documents and other image documents. The DL stack includes layers closer to an input layer pre-trained to perform image recognition before exposing a second set of layers further from the input layer to the labelled ground truth data for the image-borne sensitive documents and other image documents. Disclosed is storing parameters of the trained master DL stack for inference from production images, distributing the stack with the stored parameters to organizations and permitting them to perform update training of the DL stack using examples of the organization sensitive data in images and to save parameters of the updated DL stacks, and use respective updated DL stacks to classify production images.
Source: USPTO / EPO open patent data. Objective bibliographic and citation counts.