👁️ Part 6 of 6

Implement Computer Vision
Solutions on Azure

Computer Vision covers making machines understand visual content — images and video. This part covers Azure AI Vision (Image Analysis 4.0, OCR Read API, smart cropping), the Face API, Custom Vision for training image classifiers and object detectors, and Azure AI Video Indexer with Spatial Analysis for real-time video intelligence.

15–20%

of exam questions

topic files

~20

exam sub-skills

🔒 2

Limited Access features

your progress

💡

What This Part Covers

Computer Vision is about teaching machines to see. Azure AI Vision (Image Analysis 4.0) analyses images for tags, objects, captions, people, and text (via the Read OCR API) — no training required. Azure Custom Vision lets you train your own image classifier or object detector using your own labelled photos. Azure AI Video Indexer extracts rich insights from video: transcripts, faces, labels, scenes, and topics. Spatial Analysis uses camera feeds to detect and track people movement in physical spaces.

📋 Prerequisites — What to know before starting Part 6

✅

REST APIs and endpoints (Part 1)

Every Vision service is called via a REST API with an endpoint URL and key. You already know this pattern.

✅

Async POST-then-poll pattern (Part 5)

The OCR Read API uses this same pattern — POST image → get operation URL → poll for results.

✅

Confidence scores (Part 5)

Every Vision result comes with confidence scores on tags, objects, captions, and OCR words — same concept as Document Intelligence.

✅

Limited Access gating (Part 1 & 5)

Face API identification and the Custom Vision face features require Microsoft approval. Same policy as ID Document model — applies even more here.

Topics in this Part

Study in order — Topic 1 (Image Analysis) introduces the core Vision building blocks that Topics 2 and 3 extend.

🖼️ Analyze Images with Azure AI Vision

→

Azure AI Vision Image Analysis 4.0 — visual features (tags, objects, captions, people, brands), the OCR Read API (async text extraction from images and PDFs), handwriting recognition, smart cropping, background removal, the Face API (detection, identification, Limited Access), and Python code with every line explained.

Image Analysis 4.0 Visual Features OCR Read API Face API Smart Crop Limited Access

🎯 Custom Vision Models

→

Training your own image classifier and object detector. Image classification vs object detection, labelling with bounding boxes, Quick vs Advanced Training, evaluation metrics (Precision, Recall, mAP, probability threshold), publishing iterations, consuming via the prediction endpoint. Includes the Custom Vision retirement warning.

Classification vs Detection Bounding Boxes Precision & Recall mAP Probability Threshold Prediction API

🎥 Video Analysis & Spatial Intelligence

→

Azure AI Video Indexer — indexing videos for faces, transcripts, labels, scenes, keywords, and topics. Trial vs ARM-connected account types. Querying video insights via API. Spatial Analysis with Azure Vision — detecting and tracking people in physical spaces using camera feeds, operations (PersonCount, PersonCrossingLine, PersonZoneCrossing) and edge deployment via containers.

Video Indexer Transcripts Scenes & Shots Spatial Analysis PersonCount Edge Container

Ready to start Part 6?

Start with Topic 1 — Image Analysis is the foundation all Computer Vision builds on.

Study Flashcards Take Part Quiz Start Topic 1 →

Implement Computer VisionSolutions on Azure

Topics in this Part

Implement Computer Vision
Solutions on Azure