In this hands-on lab, you are working as an AI Engineer for The Root, a chain of supermarket stores. You’ve been asked to perform object detection of some vegetables so the company can sort out their inventory. You’ll use the Azure Custom Vision portal to upload pre-provided images, label the individual objects as either bell peppers or carrots, and then train and test your model.
Learning Objectives
Successfully complete this lab by achieving the following learning objectives:
- Log In to the Azure Portal and the Custom Vision Portal
Log in to the Azure portal using the credentials provided on the lab page. Be sure to use an incognito or private browser window to ensure you’re using the lab account rather than your own. Then, log in to the Custom Vision portal using the Azure portal credentials.
- Download the Images from the GitHub Repo
Download the Foods.zip file from the GitHub repository to your computer and extract the files locally.
- Create an Object Detection Project and Train the Model
In the Custom Vision portal, create an Object Detection project with the domain of General (A1). Upload and tag the images of bell peppers and carrots from the Training Images folder, and then train the model. Make sure to select Quick Training when training the model!
- Test It!
Use the Quick Test feature to upload images from the Testing Images folder and test if the vegetables are detected correctly in each image.
