An AI model to arm Canada against threats and disasters.


A Canadian federal government agency engaged us to develop an AI model and mobile app that could arm our country against chemical, biological, radio-logical and nuclear defence threats.

shutterstock_725365696_adjusted.jpg
 
 

Objectives

Develop an AI model that can detect, identify and characterize volatile organic compounds and will provide actionable intelligence for situation assessment, scene management, triage, disaster mitigation and eventually site restoration.

Applied Intelligence

We developed a custom AI solution to rapidly and accurately identify biological and chemical agents using colormetric sensor array (CSA) data images. The AI solution incorporates Machine Learning techniques that include:

  • IMAGE PROCESSING

  • VIRTUAL CLASSIFICATION

  • PRINCIPAL COMPONENT ANALYSIS FOR DIMENSIONALITY REDUCTION

  • MULTI-LAYER PERCEPTIONS

  • STACKED AUTO-ENCODERS

  • SUPPORT VECTOR MACHINES

  • RECURRENT NETWORKS

The mobile application of this technology is in the demonstration stage of this project and other object detection, classification and prediction tasks. The mobile app can perform tasks in a non-networked environment and synchronize with enterprise applications when network connectivity is desirable or available. The mobile app has been developed with using React Native to provide possible cross platform support.

 Augmented reality feature extraction demonstration

Augmented reality feature extraction demonstration

Results

The results and product have been extremely successful and the AI model used has already been adapted for micro-bacterial classifications, PTSD, VOC research, opioid research and agriculture use cases.



 
We developed a custom AI solution to rapidly and accurately identify biological and chemical agents.

Other Use Cases

ASSET MANAGEMENT

GEO-SPATIAL ANALYSIS

MONITORING

PTSD RESEARCH

VOC RESEARCH