Microsoft Applied Skills: Build an Azure AI Vision solution

Sign up for a free Azure account, explore the portal, and learn about the 12-month free products and 200 USD credit, while managing resource groups, subscriptions, and basic portal navigation.

Create a computer vision resource in Azure by selecting a subscription, resource group, region, and a pricing tier. Then access Vision Studio via portal.azure.com to deploy and test the resource.

Learn image analysis in Vision Studio by adding captions and dense captions to images, and read JSON results with confidence scores and bounding boxes.

Explore computer vision models for video retrieval, object detection with bounding boxes, and background removal, plus OCR and smart cropping that prioritizes people and faces.

Explore building an azure ai vision solution by creating a computer vision resource, using vision studio to detect shelf products, generate captions, tags, and smart cropped thumbnails.

Explore how to create and use a custom vision resource in Azure to train and predict models for identifying individual products, including pricing, resource setup, and uploading and tagging images.

Create and configure a Custom Vision AI project, selecting image classification with multi-label, choosing domain general A1, then upload and label images before training the model.

Train, evaluate, and consume a custom Azure AI Vision model, publish iterations, test predictions, and refine results using a probability threshold and additional training data.

Create a multi-class food classifier in Custom Vision by uploading 19 blueberry and 20 grape images, tagging them, training the model, and achieving 75% precision and 75% recall.

Explore how the ai vision service detects faces in images with Vision Studio, capturing coordinates and 27 facial points, mask detection, and attributes like glasses, pose, exposure, and occlusion.

Explore how to analyze video with Vision Studio and Azure AI Video Indexer, extracting frames, OCR, faces, celebrities, brands, topics, transcripts, and multilingual insights for searchable media.

Create a custom image analysis model in Vision Studio by setting up a blob storage container, data set, coco file, and training an image classification or object detection model.

Download and install Visual Studio Community, then create a new C# console app for AI vision analysis, choosing the dotnet desktop development workload and preparing to add a NuGet package.

Explore Visual Studio, navigate the Solution Explorer, dock panels, enable line numbers, and switch to dark mode, then install the Azure AI Vision Image Analysis NuGet package.

Learn to configure the image analysis client with an endpoint and key, analyze an image URL, and retrieve captions, dense captions, objects, OCR, and tags.

Learn to send an image to Azure Computer Vision with the image analysis client in C#, configure endpoint and key, and request features like caption and objects.

Explore the image analyzer result properties, including caption, dense captions with bounding boxes, metadata such as height, width, and model version, and ocr read blocks; learn bounding box versus polygon.

Implement the image analyzer result in C# by processing captions, dense captions, tags, objects, and OCR text with null checks.

Learn to retrieve people, objects, smart crops and read using C#, extract bounding boxes and confidence, join coordinates for display, and loop through tags, lines, and words for detailed output.

Download and install Visual Studio Code, create a C# console app, use the terminal to run dotnet commands, and add Azure AI image analysis NuGet package to run hello world.

Create a Python solution in Visual Studio Community, install the Python developer workload, and set up a Python project with the Azure AI Vision Image Analysis SDK via pip.

Use the Azure image analysis Python client with endpoint and key to extract captions, tags, objects, dense captions, and smart crops, including optional people analysis.

Pause at a breakpoint, inspect variables with a watch, drill down into results, and step through code using debugging tools for Python and C# to locate errors.

Create a custom vision prediction client in C# with a key and endpoint, then classify or detect images using project ID and iteration name, and output tag name and probability.

Install Azure Cognitive Services Custom Vision in Python, download the image via URL, connect to the endpoint with a prediction key, and run predictions using project ID and iteration name.

Learn to tidy Azure by deleting unneeded resources and resource groups, confirm deletion by name, and know that deleting a group is permanent and stops costs for all contained resources.

Complete the Azure AI Vision course by creating, analyzing, and training a custom image analysis model, then explore the Microsoft Applied Skills assessment through the interactive lab.

Complete the course by analyzing images with a custom image analysis model, generating captions, dense captions, and tags using C# and Python.