7.5. Practice Questions: Computer Vision

Name: Microsoft Azure AI Fundamentals
Price: 24.99 USD
Availability: InStock
Author: Alvin Varughese

Test your understanding of computer vision concepts and Azure AI Vision services.

Question 21: Which computer vision capability returns bounding box coordinates?

Rationale: Object detection locates objects and returns bounding boxes with coordinates. Image classification returns only a single label.

Question 22: A company needs to extract text from scanned medical records. Which capability?

Rationale: OCR extracts text from images. Since the input is scanned images (not text), this is computer vision, not NLP.

Question 23: What are the TWO specialized domain models supported by Azure AI Vision for image categorization?

Rationale: Azure AI Vision supports only celebrities and landmarks as specialized domain models—not animals, cars, or plants.

Question 24: Which feature associates images with descriptive metadata tags?

Rationale: Tagging associates images with metadata that summarizes attributes. Categorization assigns to predefined categories.

Question 25: Which output does image description include with each phrase?

Rationale: Image descriptions include confidence scores indicating certainty about each generated description.

Question 26: A farmer wants to track livestock in a field. Which capability?

Rationale: Tracking animals requires locating them in images—object detection provides positions via bounding boxes.

Question 27: Which THREE capabilities does Azure AI Vision eliminate the need for?

Rationale: Pre-built models eliminate choosing, training, and evaluating. You still must create Azure resources and call the API.

Question 28: Which Azure AI Face feature identifies if someone wears glasses?

Rationale: Face attributes include accessories like glasses, headwear, and masks.

Question 29: Which capability classifies individual pixels in an image?

Rationale: Semantic segmentation classifies every pixel, creating a detailed category map of the image.

Question 30: A quality control system needs to identify defective products on an assembly line. Which approach?

Rationale: Binary classification (defective vs. normal) using image analysis is the correct approach for visual quality control.

Written byAlvin Varughese

Founder•15 professional certifications