AI-900 Practice Question: Describe features of computer vision workloads on Azure
A warehouse deploys cameras to automatically process incoming packages. The system must read the serial numbers printed on each package label to update inventory records. The labels often have varied fonts and sizes, and may be slightly rotated. Which Azure Computer Vision capability should be used to extract the serial numbers?
⚠ Common exam trap
It's easy for candidates to confuse object detection (which can 'see' labels) with OCR, not realizing that object detection only locates objects without reading any text content on them.
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
✓
Optical Character Recognition (OCR)
Optical Character Recognition (OCR) is the correct Azure Computer Vision capability because it is specifically designed to extract printed or handwritten text from images, including serial numbers with varied fonts, sizes, and rotations. Azure's OCR API (part of Computer Vision) can handle skewed or rotated text by automatically detecting and correcting orientation before recognizing characters, making it ideal for warehouse labels that are not perfectly aligned.
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 is incorrect because it only localizes and classifies objects within a frame—for example, it can draw a bounding box around a shipping box and label it 'box' or 'pallet', but it does not read the text printed on that box. The output is a list of object classes, confidence scores, and coordinates, with no character-level extraction. Even if a detector identifies a label as a separate object, it would not return the serial number; that requires a dedicated OCR step after detection.
When this WOULD be correct
A question asks: 'Which Azure Computer Vision capability should be used to identify the locations of different types of packages in an image?' Then object detection would be correct because it provides bounding boxes around objects.
- ✓
Optical Character Recognition (OCR)
Why this is correct
Optical Character Recognition (OCR) is the correct choice because Azure Computer Vision's Read API uses deep learning models to extract printed and handwritten text from images, returning each word or line as a string alongside its bounding box and confidence score. In a warehouse scenario, OCR can accurately capture serial numbers from labels on incoming boxes, even when the label is rotated or partially obscured. This capability directly solves the business need, whereas the other computer vision tasks only produce object-level or face-level metadata and never decode alphanumeric content.
- ✗
Image classification
Why it's wrong here
Image classification is wrong because it assigns a single label or a set of class probabilities to the entire image based on the dominant visual content, such as 'cardboard box' or 'warehouse aisle', without any pixel-level localization. It cannot extract specific text characters because its output is purely categorical and does not include any string decoding. This means the camera could tell you that a box is present, but it could never tell you which serial number is on that box, so it fails to meet the requirement.
When this WOULD be correct
A question asks: 'Which Azure Computer Vision capability should be used to categorize images of warehouse packages into predefined categories such as box, envelope, or pallet?'
- ✗
Facial recognition
Why it's wrong here
Facial recognition is irrelevant here because its purpose is to detect, analyze, and match human faces by comparing facial landmarks and encoding them into vectors—it has no mechanism for interpreting alphanumeric codes on warehouse labels. Azure Face API is designed for identity verification, emotion detection, and similar demographic analysis, not for reading text. Applying facial recognition to cartons would either fail to detect a face or return an error, and it would never output a serial number.
When this WOULD be correct
A question asks: 'A security system must identify employees entering a restricted area by analyzing camera feeds. Which Azure Computer Vision capability should be used?' In that scenario, facial recognition would be correct.
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.
✓Optical Character Recognition (OCR)Correct answer▾
Why this is correct
Optical Character Recognition (OCR) is the correct choice because Azure Computer Vision's Read API uses deep learning models to extract printed and handwritten text from images, returning each word or line as a string alongside its bounding box and confidence score. In a warehouse scenario, OCR can accurately capture serial numbers from labels on incoming boxes, even when the label is rotated or partially obscured. This capability directly solves the business need, whereas the other computer vision tasks only produce object-level or face-level metadata and never decode alphanumeric content.
✗Object detectionWrong answer — click to see why▾
Why this is wrong here
Object detection identifies and locates objects within an image but does not extract text content. The requirement is to read serial numbers, which requires text extraction, not object localization.
★ When this WOULD be the correct answer
A question asks: 'Which Azure Computer Vision capability should be used to identify the locations of different types of packages in an image?' Then object detection would be correct because it provides bounding boxes around objects.
Why candidates choose this
Candidates may confuse object detection with OCR because both involve analyzing images, but object detection focuses on finding objects, not reading text.
✗Image classificationWrong answer — click to see why▾
Why this is wrong here
Image classification assigns a single label to an entire image (e.g., 'package' or 'label'), but cannot extract specific text like serial numbers from varied fonts and rotations.
★ When this WOULD be the correct answer
A question asks: 'Which Azure Computer Vision capability should be used to categorize images of warehouse packages into predefined categories such as box, envelope, or pallet?'
Why candidates choose this
Candidates may confuse 'classifying' the content of an image with 'reading' text, assuming that identifying the label type is sufficient to extract its data.
✗Facial recognitionWrong answer — click to see why▾
Why this is wrong here
Facial recognition is designed to identify or verify individuals by analyzing facial features, not to extract text from images. The question requires reading serial numbers, which is a text extraction task, not a face-related task.
★ When this WOULD be the correct answer
A question asks: 'A security system must identify employees entering a restricted area by analyzing camera feeds. Which Azure Computer Vision capability should be used?' In that scenario, facial recognition would be correct.
Why candidates choose this
Candidates may confuse 'recognition' in facial recognition with general pattern recognition, mistakenly thinking it can recognize any pattern including text, or they may not distinguish between different Computer Vision APIs.
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
Azure Machine Learning Studio
Key term
Computer vision
Computer vision is a field of artificial intelligence that enables computers to interpret and make decisions based on visual data from the world, such as images and videos.
Key term
Inventory
Inventory is the complete list of hardware, software, and network assets that an organization owns or manages, tracked for maintenance, security, and budgeting purposes.
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.