AWS DeepLens
Developer Guide

Build and Run the Head Pose Detection Project with TensorFlow-Trained Model

In this tutorial, we will walk you though an end-to-end process to build and run an AWS DeepLens project to detect head poses. The project is based on the Head pose detection sample project. Specifically, you'll learn:

  1. How to train a computer vision model using the TensorFlow framework in Amazon SageMaker, including converting the model artifact to the protobuff format for use by AWS DeepLens.

  2. How to create an inference Lambda function to infer head poses based on frames captured off the video feed from the AWS DeepLens video camera.

  3. How to create an AWS DeepLens project to include the Amazon SageMaker-trained deep learning model and the inference Lambda function.

  4. How to deploy the project to your AWS DeepLens device to run the project and verify the results.