Description:
Part 1: Data Augmentation — Training Better Image Classification Models with Less Data

Image classification can be tricky, especially as objects may appear under different lighting conditions and in different locations, orientations or zoom levels. While capturing new, original data is the ideal approach for creating a robust dataset for machine learning model training, we can use data augmentation to automatically generate data when original data capture is overly difficult or time consuming. In this workshop, we will use Google Colab and Edge Impulse to examine various image augmentation techniques to generate new data from a small initial dataset.

Part 2: Develop and Deploy Vision Models at the Edge – with No Code

In this workshop, attendees will learn how to quickly and easily create vision models and deploy them at the edge without writing code. We’ll work hands-on with the Edge Impulse cloud-based development environment to develop a model, and then deploy it on hardware using a Texas Instruments SK-TDA4VM starter kit. The SK-TDA4VM starter kit is based on TI’s TDA4VM processor, which features 8 TOPS of deep learning performance and low power consumption. This combination of development techniques, tools and hardware is ideally suited to a wide range of vision applications, including industrial, agriculture and security applications.

After the workshop, join your Edge Impulse hosts for food and drinks at a complimentary reception on the outdoor terrace!

Session: 3:00 – 6:00 pm Reception: 6:00 – 7:30 pm