How to Build a Computer Vision App using Workflows and Custom Python Blocks

This video guides you through the process of building a computer vision application to detect hand gestures, add specific data to the output (e.g., attach timestamps when detections last occurred), or overlay customized labels and visualizations. To accomplish this, we use the Workflows editor with a combination of default tools and custom Python blocks.

About the tools in this video:

The Workflows editor in Roboflow lets you visualize and build automations that execute tasks depending on the input, combine vision models and third-party services to build powerful vision AI applications, and do more.

With the addition of the new Custom Block in Workflows, you can add custom Python code to your automations. This allows you to create more complex and flexible processes when deploying your computer vision model.

What you’ll see in this video:
- Introduction to the Workflows editor and its library of ready-made blocks
- Step-by-step guide to adding a Custom Block for Python code
- How to run Workflows with Custom Blocks in local edge environments

Helpful resources:

Talk to us

[MathaGoram]

With such a comprehensible presentation, it would be our loss not to dip our toes in the water. Just completed the classic NVIDIA Jetson tutorial (on Orin), but this one is much better!