Pitfalls of Emotion Detection in Production – Justin Shenk, Peltarion

Lessons learned with solving facial expression recognition use cases at Intel and with a European consultancy. Introduction of deep neural network model training and tooling. Overview of deployment infrastructure and choices.

Key Takeaways

  • Don’t reinvent the wheel with data models
  • Strike a balance between test accuracy and data throughput
  • Choose the most sustainable tooling with the high performance

Add comment