Curriculum Overview: Azure Computer Vision Workloads

This unit explores the fundamental concepts and Azure services associated with Computer Vision. Representing 15–20% of the AI-900 exam, it focuses on how machines can interpret visual data and the specific Microsoft Azure tools available to implement these solutions.

Prerequisites

Before beginning this module, learners should have a foundational understanding of the following:

• Basic AI Concepts: Familiarity with what Artificial Intelligence is and its role in modern technology.
• Azure Fundamentals: A high-level understanding of cloud computing and the Microsoft Azure ecosystem.
• Machine Learning Principles: Knowledge of "features" and "labels," and the difference between training and validation datasets (typically covered in Unit 2).
• Responsible AI: Understanding the six guiding principles (Fairness, Reliability, etc.) as they apply specifically to sensitive visual tasks like facial recognition.

Module Breakdown

Learning Objectives per Module

1. Identify Common Types of Computer Vision Solutions

• Image Classification: Distinguish between different classes of images (e.g., "Is this a cat or a dog?").
• Object Detection: Identify the location and type of multiple objects within an image using bounding boxes.
• Optical Character Recognition (OCR): Extract printed or handwritten text from images and documents.
• Facial Detection & Analysis: Detect human faces, identify attributes (age, emotion), and recognize individual identities.

2. Identify Azure Tools and Services

• Azure AI Vision: Understand its versatility in analyzing images, video, and extracting spatial metadata.
• Azure AI Face: Recognize when to use specialized facial recognition vs. general object detection.

Success Metrics

To demonstrate mastery of this curriculum unit, learners must be able to:

• Differentiate between Image Classification (what is in the photo) and Object Detection (where is it in the photo).
• Select the appropriate Azure service (Vision vs. Face) for a specific business case.
• Explain how OCR converts pixel data into machine-readable text characters.
• Identify the specific technical attributes returned by Facial Analysis (e.g., head pose, smile detection).

Real-World Application

Computer Vision is not just theoretical; it powers critical infrastructure and consumer technology. Below is a conceptual representation of Object Detection in a self-driving car or security scenario.

Practical Scenarios:

• Retail: Using Object Detection to monitor shelf stock levels automatically.
• Healthcare: Applying Image Classification to X-rays to flag potential anomalies for radiologist review.
• Finance: Using OCR to automate the processing of physical invoices and receipts into accounting software.
• Security: Utilizing Azure AI Face for secure building access or identity verification in mobile apps.