Computer Vision Fundamentals with Google Cloud
Google Skills
A practical course on solving image problems on Google Cloud, working up through three levels of effort. It starts with the pre-built Vision API, which labels images, detects faces and landmarks, and pulls text out of pictures with no training involved. It moves to AutoML Vision on Vertex AI, where you supply labeled examples and the platform trains a classifier for you — the worked example identifies damaged car parts. Then it builds models from scratch in Python and tf.keras: linear, then deep neural, then convolutional, with the mechanics of convolutions, pooling layers, and their parameters explained along the way. The final part deals with what ends most image projects — too little training data — through preprocessing, data augmentation, and transfer learning, and with raising accuracy without overfitting. About eight hours. The concept content is free; the hands-on labs, which are a required part of this course, run in the Google Cloud console and need a Google Skills subscription or credits.
What you'll learn
- The kinds of problems computer vision solves, and where each shows up in practice
- Labeling images and detecting faces, landmarks, and text with the pre-built Vision API
- Training a custom image classifier with AutoML Vision on Vertex AI
- Building your own classifiers in tf.keras: linear, deep neural, and convolutional
- What convolutions and pooling layers do, and the parameters that control them
- Preprocessing images, and handling scarce training data with augmentation and transfer learning
- Raising accuracy while avoiding overfitting
Frequently asked questions about Computer Vision Fundamentals with Google Cloud
Who is Computer Vision Fundamentals with Google Cloud for?
Developers and data professionals working on Google Cloud who want to choose between a pre-built vision API, AutoML, and a custom model — and build all three.
What are the prerequisites for Computer Vision Fundamentals with Google Cloud?
None stated. Some Python will help all the same: the custom-model sections build image classifiers with the tf.keras API, and the hands-on labs run in a live Google Cloud environment provisioned through Google Skills rather than on your own machine.
Do you need to code for Computer Vision Fundamentals with Google Cloud?
Yes — Computer Vision Fundamentals with Google Cloud involves hands-on coding.
Why we suggest this course
A well-sequenced tour of the real choices in an image project — call a pre-built API, train an AutoML model on your own labeled data, or write the network yourself — with labs for each, so the trade-off between effort and control is something you experience rather than read about. One thing to know: the labs here are required rather than optional, and they run in a live Google Cloud environment that needs a Google Skills subscription or credits; finishing the course earns a Google skill badge.