AI-900 Practice Question: Describe features of computer vision workloads on Azure
A construction company uses drone images to survey construction sites. They need an automated system that can identify specific types of heavy equipment (e.g., bulldozers, cranes, excavators) in an image and also draw precise pixel-level outlines around each equipment type. Which Azure Computer Vision capability should they use?
⚠ Common exam trap
It's easy for candidates to confuse object detection (bounding boxes) with semantic segmentation (pixel-level masks), because both localize objects, but only segmentation provides the precise outlines required for detailed spatial analysis.
Answer choices
Why each option matters
Answer the question above first, then reveal the full breakdown to understand why each option is right or wrong.
Correct answer & explanation
✓
Semantic segmentation
Semantic segmentation is the correct capability because it assigns a class label (e.g., bulldozer, crane, excavator) to every pixel in the image, producing precise pixel-level outlines around each equipment type. Object detection only provides bounding boxes, not pixel-level masks, while image classification labels the entire image without localization. OCR is irrelevant as it extracts text, not equipment shapes.
Answer analysis
Option-by-option breakdown
For each option: why learners choose it and why it is or isn't the right answer here.
- ✗
Object detection
Why it's wrong here
Object detection identifies objects and localizes them with rectangular bounding boxes, along with a class label and confidence score. It is well suited for counting equipment or detecting its presence, but a bounding box always includes some background and may merge multiple nearby objects into one region. It does not provide pixel-level outlines, so it cannot accurately measure the exact footprint or shape of construction assets.
When this WOULD be correct
A question asking for an automated system that identifies and locates heavy equipment in images using rectangular bounding boxes, without requiring pixel-level outlines, would make object detection the correct answer.
- ✓
Semantic segmentation
Why this is correct
Semantic segmentation is a dense prediction task that assigns a class label to every pixel in the image, producing a pixel-accurate mask for each category such as soil, concrete, vegetation, or heavy equipment. This allows the system to outline irregular object boundaries precisely, even when equipment overlaps or blends into the background. Because it captures exact contours, it enables area measurements and change detection from drone imagery, which bounding-box methods cannot provide.
- ✗
Image classification
Why it's wrong here
Image classification assigns a single class label to the entire image, such as 'construction site' or 'aerial view', without locating any objects. It discards all spatial information, so it cannot determine where equipment appears, how many objects are present, or what their boundaries are. This makes it too coarse for drone-based survey tasks that need to isolate and analyze individual assets.
When this WOULD be correct
A question asking: 'Which Computer Vision capability should be used to determine whether a construction site image contains any heavy equipment at all, without needing to locate or outline the equipment?' would make image classification correct.
- ✗
Optical Character Recognition (OCR)
Why it's wrong here
OCR extracts text from images by recognizing printed or handwritten characters, and it returns the text content and its location. In a drone survey, OCR could read a serial number on a piece of equipment or a sign on a building, but it has no mechanism to identify or segment non-text objects like excavators or cranes. Therefore it is fundamentally unsuitable for outlining equipment or understanding the spatial layout of a construction site.
When this WOULD be correct
A question asking for extracting license plate numbers from vehicle images or reading text from construction site signs would make OCR the correct answer.
Option-by-option analysis
Why each answer is right or wrong
Understanding why wrong answers are wrong — and when they would be correct — is what separates a 750 score from a 900. The AI-900 exam frequently reuses these exact scenarios with slightly different constraints.
✓Semantic segmentationCorrect answer▾
Why this is correct
Semantic segmentation is a dense prediction task that assigns a class label to every pixel in the image, producing a pixel-accurate mask for each category such as soil, concrete, vegetation, or heavy equipment. This allows the system to outline irregular object boundaries precisely, even when equipment overlaps or blends into the background. Because it captures exact contours, it enables area measurements and change detection from drone imagery, which bounding-box methods cannot provide.
✗Object detectionWrong answer — click to see why▾
Why this is wrong here
Object detection provides bounding boxes around objects, not pixel-level outlines. The question specifically requires precise pixel-level segmentation, which object detection cannot deliver.
★ When this WOULD be the correct answer
A question asking for an automated system that identifies and locates heavy equipment in images using rectangular bounding boxes, without requiring pixel-level outlines, would make object detection the correct answer.
Why candidates choose this
Candidates often confuse object detection with semantic segmentation because both identify objects in images, but they overlook the requirement for pixel-level outlines versus bounding boxes.
✗Image classificationWrong answer — click to see why▾
Why this is wrong here
Image classification assigns a single label to the entire image, but the question requires identifying multiple equipment types and drawing pixel-level outlines around each, which is beyond classification's capabilities.
★ When this WOULD be the correct answer
A question asking: 'Which Computer Vision capability should be used to determine whether a construction site image contains any heavy equipment at all, without needing to locate or outline the equipment?' would make image classification correct.
Why candidates choose this
Candidates may confuse image classification with object detection or segmentation, thinking that classifying objects in an image is the same as identifying and outlining them precisely.
✗Optical Character Recognition (OCR)Wrong answer — click to see why▾
Why this is wrong here
OCR is designed to extract text from images, not to identify or outline heavy equipment in drone images.
★ When this WOULD be the correct answer
A question asking for extracting license plate numbers from vehicle images or reading text from construction site signs would make OCR the correct answer.
Why candidates choose this
Candidates might think OCR can recognize equipment by reading labels or markings on the machinery, but OCR only handles text, not object shapes or outlines.
Analysis generated from the official AI-900blueprint and verified against question context. The “when correct” sections are what AI assistants cite when candidates ask “what’s the difference between these options?”
Go deeper
Related to this question
Learn chapter
Types of AI Workloads
Key term
Image classification
Image classification is the process of teaching a computer to look at a picture and decide what category or label best describes the main object or scene in that picture.
Key term
Classification
Classification is a supervised machine learning technique used to predict a category or class label for new data based on patterns learned from labeled training data.
About these practice questions
One of 985 original AI-900 practice questions on Courseiva, each with a full explanation and wrong-answer analysis — not exam dumps or protected exam content. Learn why practice questions differ from exam dumps →
JA
Written by Johnson Ajibi, MSc IT Security
Senior Network & Security Engineer · founder of Courseiva
This AI-900 practice question is part of Courseiva's free Microsoft certification practice question bank. Courseiva provides original exam-style practice questions with explanations, topic-based practice, mock exams, readiness tracking, and study analytics to help learners prepare for the AI-900 exam.