1. What This Means in the Exam
Using Azure AI Foundry, you should understand how to:
Analyse images (vision)
Extract information from images
Generate new images from prompts
Integrate these capabilities into applications
👉 Think:
“See images vs create images”
🖼️ 2. Computer Vision (Image Understanding)
đź“Ś What it does
Processes existing images to extract insights
🔑 Capabilities
Image classification → What is in the image
Object detection → Where objects are
OCR (optical character recognition) → Read text from images
Face detection → Identify faces (not identity in AI-901)
âś… Example Scenarios
Detect defects in manufacturing → Object detection
Extract text from invoices → OCR
Categorise photos → Image classification
🎨 3. Image Generation (Create Images)
đź“Ś What it does
Creates new images from text prompts
🔑 Capabilities
Generate images from descriptions
Create variations of images
Edit images (inpainting, style changes)
âś… Example Scenarios
Generate marketing visuals
Create product mockups
Produce illustrations from descriptions
🔄 4. How It’s Implemented in Foundry
Using Azure AI Foundry:
For Computer Vision:
Select a vision model
Send image input
Receive structured output (labels, text, objects)
For Image Generation:
Select an image generation model
Provide a text prompt
Generate image
Refine prompt if needed
⚖️ 5. Key Exam Distinctions
CapabilityComputer VisionImage GenerationInputImageTextOutputInsights/dataNew imagePurposeUnderstandCreate
⚠️ 6. Common Exam Traps
❌ Confusing OCR with image generation
✔️ OCR = extract text
✔️ Generation = create image
❌ Thinking vision = only classification
✔️ Includes detection, OCR, faces
❌ Forgetting prompt role in image generation
✔️ Prompt = key input
🧠7. Simple Memory Trick
“See vs Create”
See → Computer Vision
Create → Image Generation
🎯 8. Example Exam Scenario
Scenario:
“An app must read invoice details from scanned documents”
✔️ Use:
Computer Vision (OCR)
Scenario:
“A marketing team wants to generate product images from descriptions”
✔️ Use:
Image generation
âś… Summary
To implement AI solutions with vision and image generation using Foundry:
Use Azure AI Foundry
For vision: Analyse images (classification, OCR, detection)
For generation: Create images from prompts
Match: Input type → capability → model
If you want, I can give you a full AI-901 mock exam covering ALL workloads (text, speech, vision, agents)—that’s the fastest way to lock everything in before the exam.