Solana Networks wins contract to provide Intelligent Data Annotation solution for AI Modeling

Ottawa, ON – 
Jun 10, 2020

Defence R&D Canada (DRDC) has awarded Solana Networks a contract to develop an intelligent data annotation solution for building AI models.

In recent years, AI-based solutions to detect and classify vehicles/objects in video and image have witnessed significant advances in accuracy. However, a key challenge in building accurate AI models is the lack of sufficiently large and accurately labelled/annotated datasets. Current manual methods for AI model dataset labeling are cumbersome, time consuming and not viable for building large scale datasets.

The team is developing a human-in-the-loop semi-automated AI annotation software tool. The tool will combine analyst intelligence with a deep learning image and video processing system for object recognition and tracking using convolution neural networks. The labeled datasets are to be used for training AI algorithms to perform vehicle detection and classification, and develop human pose estimation and activity patterns recognition models.

Sufficiently large and accurate labelled datasets are a key requirement for building AI models in different domains. Solana has developed a number of solutions which assist AI practitioners to rapidly build labeled datasets for AI model building in different domains including cyber security, network traffic classification and object detection.