1. What This Means in the Exam
Using Azure AI Foundry, you should recognise how to:
Take an image as input
Send it to a vision model/service
Receive insights about the image
Display the results in a simple app
👉 Think:
“Send image → analyse → show result”
🧩 2. What is a Lightweight Application?
A lightweight app is a simple client, such as:
Console app
Basic web app
Small script
It:
Accepts an image
Calls a vision service
Displays the output
🖼️ 3. Vision Capabilities You Might Use
📌 Common capabilities (AI-901 focus)
🔍 Image Classification
Identify what is in the image
📦 Object Detection
Locate objects within the image
📝 OCR (Optical Character Recognition)
Extract text from images
🙂 Face Detection
Detect presence of faces
🔄 4. Typical Architecture (Exam-Friendly)
User uploads image
↓
Lightweight app
↓
Vision model (in Foundry)
↓
Structured output (labels/text/objects)
↓
App displays result
⚙️ 5. High-Level Implementation Steps
1️⃣ Get Image Input
Upload file or provide image URL
2️⃣ Call Vision Service
Using Azure AI Foundry:
Send image to model
3️⃣ Receive Results
Examples:
Labels: “car”, “person”
Extracted text: invoice details
Detected objects: bounding boxes
4️⃣ Display Output
Show results in UI or console
🎯 6. Example Scenario (Exam Style)
Scenario:
“Build an application that extracts text from scanned receipts”
✔️ Use:
Lightweight app
OCR capability
Display extracted text
Scenario:
“Identify objects in images uploaded by users”
✔️ Use:
Object detection
⚖️ 7. Key Exam Distinctions
| Feature | Lightweight App | Vision Model |
|---|---|---|
| Role | Sends/receives data | Performs analysis |
| Contains AI logic | ❌ No | ✅ Yes |
| Built with SDK/API | ✅ Yes | Used via endpoint |
⚠️ 8. Common Exam Traps
❌ Thinking you must train a model
✔️ Use prebuilt vision capabilities
❌ Confusing with image generation
✔️ Vision = analyse, not create
❌ Overcomplicating the solution
✔️ It’s just a simple client calling a service
🧠 9. Simple Memory Model
“Image → Analyse → Insight”
Image → input
Analyse → vision model
Insight → result displayed
✅ Summary
To build a lightweight application with vision capabilities in AI-901:
Use Azure AI Foundry
Create a simple app (console/web)
Send image to vision service
Receive structured insights
Display results