We help the Navy and DoD analyze and act on the flood of visual data they now collect from camera-mounted field equipment and
personnel. To do so, we offer computer vision solutions that use machine learning to automatically produce actionable insights from
large-scale visual data.
What We Do:
COMPUTER VISION DRIVEN BY SYNTHETIC DATA AND MACHINE LEARNING
Agencies collect an increasing volume of visual data from cameras mounted to vehicles, drones, IoT equipment, staff members, and
more. Yet they struggle to manually process this large volume of data into usable information.
Overlab solves this problem. We develop computer vision solutions that capture visual data from images, videos and 3D
representations, and apply machine learning to process this data into actionable insights. Our solutions leverage synthetic data
capabilities that automatically trains neural networks without having to collect and hand-label large datasets of images.
— saving substantial time and effort.
COMPUTER VISION SOLUTIONS INCLUDE:
State of the art
Synthetic data
Automated Testing
Temporal Detection
State of the art
SECTORS AND USE CASES
Image Recognition & Automated Captions
We have developed computer vision solutions for both private sector and public sector entities, including the Navy and DoD. Past and
future use cases include:
Industrial Manufacturing
- Industrial manufacturing and maintenance
In-field Equipment
- In-field equipment maintenance (PMA-260)
Automated Training
- Equipment maintenance training (PMA-205)
Equipment Diagnostics
- Equipment diagnostics and fault detection
AR-GUIDED EQUIPMENT STATE RECOGNITION
The Navy and Marine Corps are pursuing Augmented Reality (AR) to create a “GPS for equipment maintenance” that will drive faster, more accurate repairs in the field.
To support them, we are developing computer vision solutions that combine 3D CAD models with AR sensor data to deliver equipment state recognition capabilities like Pose Estimation, Assembly Recognition, and Articulation Recognition.
Our solutions use high-fidelity synthetic data and leverage machine learning to adapt to new equipment and to accurately estimate state without physical markers on the equipment itself.
