In Class Object Detection with Real Time Inference on Kubernetes Deployment

Used ~20 images from class to build a simple object detector with real time inference on a flask web service deployment through Kubernetes

Actual images excluded due to privacy reasons

Training

• Use the environment.yml file to set up using Anaconda
• Annotate your images with any tool of your choice that outputs to the PASCAL-VOC XML format
• Run the code and use the model that performs well.

Inference

• Use image-recognition.yaml to deploy on Kubernetes
• Create a service to make it accessible
• Use the server endpoint to access the Flask application