Computer Vision Technologies R&D | SRI International

Toggle Menu

Computer Vision Technologies R&D

The science and technology of systems that see and understand—computer vision—offers solutions to data acquisition and exploitation challenges in defense, security, robotics, training, and the medical marketplace. Cameras and sensors have become ubiquitous: extracting meaningful information from acquired images in real time and in real world applications requires the right set of tools.

End-to-end vision capabilities include enhanced sensing, front-end video processing, and high-level analysis. An important application area is event detection and recognition to identify and extract highlights from long-form video, and to label events throughout a video so they are searchable (e.g., at what point did the candles get blown out in a birthday video?).

To meet rigorous client requirements, SRI develops, evaluates, and transitions computer vision technologies in these areas:

Core Technologies



Automated Fish Classification

FLASK provides a fast annotation and machine learning based classification capability to tackle the challenges users face that have little to no annotation data available to train a neural network.

Artificial Intelligence System that Continually Learns

SRI is developing a next-generation artificial intelligence system that applies continuous learning to become better and more reliable at performing new tasks.

Spotting Audio-Visual Inconsistencies (SAVI)

SRI is finding new ways to detect altered and tampered video.

soldier viewing video on a tablet

Adaptable Sensor System for Real-Time Data on the Front Line

DARPA taps SRI to deliver usable video and analytics to U.S. Armed Forces when and where they need it.