AI-900 Practice Question: Describe features of computer vision workloads on Azure
A social media platform wants to automatically generate alternative text descriptions for images posted by users to improve accessibility for visually impaired users. Which Azure Computer Vision capability should be used?
⚠ Common exam trap
Test-takers frequently confuse Object Detection (which only lists objects) with Image Captioning (which generates a full description), leading them to choose C because they think identifying objects is sufficient for accessibility, but screen readers need natural language descriptions, not just object labels.
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
✓
Image Captioning
Image Captioning is the correct capability because it generates human-readable descriptions of image content, which directly meets the requirement to produce alternative text for accessibility. Unlike other options, it synthesizes a complete sentence describing the scene, objects, and actions, making it ideal for screen readers.
Answer analysis
Option-by-option breakdown
For each option: why learners choose it and why it is or isn't the right answer here.
- ✗
Optical Character Recognition (OCR)
Why it's wrong here
OCR (Optical Character Recognition) uses pattern recognition algorithms to localize and transcribe printed or handwritten text into a machine-readable string. However, it operates purely on character shapes and cannot understand the visual scene, objects, or spatial relationships. As a result, OCR would only extract any text visible in the image, not generate a descriptive alternative-text caption for the entire image.
- ✓
Image Captioning
Why this is correct
Image Captioning is a multimodal AI task that combines computer vision and natural language generation to produce a grammatical, human-like description of an image's contents. It typically uses an encoder-decoder architecture, such as a CNN to extract visual features and a transformer or RNN to decode those features into a coherent caption. This makes it specifically designed to create alt-text that conveys the scene, actions, and relationships for accessibility.
- ✗
Object Detection
Why it's wrong here
Object Detection goes beyond classification by drawing bounding boxes around each detected object instance and assigning a label, such as 'dog' or 'car'. While it provides spatial locations and individual object categories, it cannot compose a holistic sentence that describes the overall scene, the interactions between objects, or the context. Therefore, it would return a list of detections rather than the fluent, comprehensive alt-text required for automatically generated image descriptions.
- ✗
Face Detection
Why it's wrong here
Face Detection is a specialized vision task that locates human faces in an image, often returning face rectangles and optional attributes like emotion, age, or gaze direction. It completely ignores all non-face content, such as buildings, animals, or actions, so its output cannot describe the broader scene for screen readers. This narrow focus makes face detection inadequate as a general-purpose alt-text generator, which needs to cover every visually important element in the image.
Go deeper
Related to this question
About these practice questions
Courseiva writes every AI-900 question from scratch — 985 in total, each with an explanation and a wrong-answer breakdown. None are copied from real exams or dumps. 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.