Extract Vision

FOUNDATION AI

Configurable machine learning, neural network, and deep learning models to extract structured data from images, videos and documents. Assemble with Language and Learning models to deploy customized end-to-end solutions.

Image Classification, Analysis and Anomaly Detection

Video Analytics, Search, and Object Tracking

Intelligent OCR, Document Classification, and Text Extraction

Intelligent OCR

Extract text from documents and media

Extract can identify and capture business critical content from hundreds of documents out of the box, including legal pleadings, medical records, common insurance documents like ACORD forms, and IRS documents like 1040s and W-2s.

The Extract AI engine can be easily configured to identify and extract text from virtually any document, including unstructured documents with inconsistent layouts.

Capture people, things, and events across video assets

Extract Vision uses transfer learning to quickly identify new types of objects with limited training data. Once configured, Extract can count and track multiple disparate objects moving in multiple directions through a video frame--like products on a factory floor, people walking through a shopping center, or cars on the road.

Foundation AI has deployed these models to help municipalities increase urban mobility and traffic safety by tracking the number, types, and direction of vehicles moving through busy intersections.

Video Object Tracking

Facial Recognition

Identify people in real-time, with limited training data

Foundation Extract facial recognition leverages state-of-the-art models to recognize people with limited training data. Extract can recognize faces in digital and analog footage, regardless of angle of capture, and even in suboptimal conditions like when they are wearing face masks.

Extract facial recognition can process archived images and video, or deployed in real time across a wide range of use-cases like forensic analysis, e-discovery, medication adherence monitoring, physical premises monitoring, and bio-metric access control.

Detect and Diagnose Medical Conditions

Not only has Foundation AI’s data science team published several papers on deep learning models to detect and diagnose diseases like tuberculosis and diabetic retinopathy, but it has also put these models in production to optimize workflows in the real world.

Foundation AI works with forward-thinking provider groups and healthcare technology companies around the world to integrate deep learning powered diagnostic imaging technology into their offerings.