Description

AI is only as good as its training data; a problem for Computer Vision where sourcing and prepping image data is slow and expensive. This is especially true for mixed reality applications as the industry transitions from AR based targets to true environmental interaction. In the near future, 5G is going to catalyze computer vision in ways that will be hugely enabling for the industry. LexSet was one of the winners of Verizon’s Built on 5G Challenge (as well as the winner for the Auggie for Startup to Watch at AWE 2019), their President and Co-Founder Leslie Karpas will unpack the problems with training computer vision models today, and explain how using synthetic data can solve them; as well at taking a look at how this dovetails with the incoming 5G future. Using the lense of LexSet’s flagship TDaaS (Training Data as a Service), we’ll go in-depth on how using 3D simulation to create photo-realistic synthetic data improves Vision AI model's overall accuracy at object recognition and spatial navigation.

Speakers

Leslie Karpas

President , LexSet