The answer is Azure Custom Vision’s object detection capability. This is correct because the retail company needs to detect and count specific product categories like Brand A and Brand B cereal from shelf images, which requires training a custom model on a labeled dataset. Unlike pre-built Computer Vision features, Custom Vision object detection allows you to upload your own images, train the model to recognize multiple product classes, and return bounding boxes with per-class counts—exactly what’s needed for inventory monitoring. On the AI-900 exam, this question tests your understanding of when to choose Custom Vision over pre-built services like Image Analysis; a common trap is selecting “image classification,” but that only labels the entire image, not individual objects. Remember: if you need to locate and count multiple instances of specific items, think “Custom object detection.” Memory tip: “Custom counts—classification can’t.”
AI-900 Practice Question: Describe features of computer vision workloads on Azure
This AI-900 practice question tests your understanding of describe features of computer vision workloads on azure. Match the stated requirement to the specific cloud service, access model, or configuration option — many options are valid in isolation but not for this scenario. A key principle to apply: custom object detection trains models to identify specific objects in images.. Once you have made your selection, read the full explanation to reinforce the concept and understand why each distractor is designed to mislead on exam day.
A retail company wants to use Azure Computer Vision to monitor shelf inventory. They need to detect whether specific products (e.g., 'Brand A cereal', 'Brand B cereal') are present on a shelf and count the number of units of each product. They have a labeled dataset with images of each product category. Which Azure Computer Vision capability should they use?
With custom object detection, you can train a model on labeled images to detect specific products (e.g., Brand A cereal) and count their occurrences, meeting the requirement.
B
Optical character recognition (OCR)
Why wrong: OCR reads text from images (e.g., product labels), but it does not identify or count the physical product objects themselves.
C
Prebuilt image analysis (Describe Image)
Why wrong: Prebuilt image analysis generates tags or captions for common objects, but it cannot be customized to detect and count specific branded products reliably.
D
Facial recognition
Why wrong: Facial recognition identifies human faces; it is not suitable for detecting retail products on shelves.
Answer the question above first, then reveal the full breakdown to understand why each option is right or wrong.
Correct answer & explanation
✓
Custom object detection
Custom object detection (A) is correct because the retail company needs to detect and count specific product categories (e.g., 'Brand A cereal', 'Brand B cereal') from images, which requires training a model on a labeled dataset of those products. Azure Custom Vision's object detection capability allows you to upload labeled images, train a model to identify and locate multiple instances of each product in an image, and return bounding boxes with counts per class. This is the only option that supports custom, multi-class object detection and counting from user-provided training data.
Key principle: Custom object detection trains models to identify specific objects in images.
Answer analysis
Option-by-option breakdown
For each option: why learners choose it and why it is or isn't the right answer here.
✓
Custom object detection
Why this is correct
With custom object detection, you can train a model on labeled images to detect specific products (e.g., Brand A cereal) and count their occurrences, meeting the requirement.
Related concept
Custom object detection trains models to identify specific objects in images.
✗
Optical character recognition (OCR)
Why it's wrong here
OCR reads text from images (e.g., product labels), but it does not identify or count the physical product objects themselves.
✗
Prebuilt image analysis (Describe Image)
Why it's wrong here
Prebuilt image analysis generates tags or captions for common objects, but it cannot be customized to detect and count specific branded products reliably.
✗
Facial recognition
Why it's wrong here
Facial recognition identifies human faces; it is not suitable for detecting retail products on shelves.
Common exam traps
Common exam trap: answer the scenario, not the keyword
The trap here is that candidates confuse prebuilt image analysis (which can 'describe' a scene) with custom object detection, not realizing that prebuilt models cannot be trained on specific product categories or provide per-object counts.
Detailed technical explanation
How to think about this question
Under the hood, Azure Custom Vision object detection uses a transfer learning approach with a ResNet-based backbone (e.g., ResNet50) fine-tuned on the user's labeled dataset, where each image includes bounding box annotations for each product instance. The model outputs a list of detected objects with confidence scores and bounding box coordinates, enabling counting by tallying detections per class. In a real-world scenario, the company would need to ensure diverse training images (varying lighting, angles, occlusions) to avoid overfitting and achieve robust shelf-monitoring accuracy.
KKey Concepts to Remember
Custom object detection trains models to identify specific objects in images.
It requires a labeled dataset with bounding boxes around target objects.
It can detect multiple instances of an object and provide their count.
It is ideal for domain-specific object recognition tasks like product identification.
TExam Day Tips
→Watch for words such as best, first, most likely and least administrative effort.
→Review why wrong options are wrong, not only why the correct option is correct.
Key takeaway
Custom object detection trains models to identify specific objects in images.
Real-world example
How this comes up in practice
A media company stores terabytes of video archives that are accessed once a year for audit purposes. Moving these objects to a cold storage tier (Azure Archive, S3 Glacier, or Google Nearline) costs a fraction of hot storage. Questions like this test whether you understand storage tiers, access frequency tradeoffs, and retrieval latency requirements.
Related glossary terms
Concepts from this question explained
These glossary pages explain the core terms tested in this AI-900 question in full detail.
Review custom object detection trains models to identify specific objects in images., then practise related AI-900 questions on the same topic to reinforce the concept.
Describe features of computer vision workloads on Azure — This question tests Describe features of computer vision workloads on Azure — Custom object detection trains models to identify specific objects in images..
What is the correct answer to this question?
The correct answer is: Custom object detection — Custom object detection (A) is correct because the retail company needs to detect and count specific product categories (e.g., 'Brand A cereal', 'Brand B cereal') from images, which requires training a model on a labeled dataset of those products. Azure Custom Vision's object detection capability allows you to upload labeled images, train a model to identify and locate multiple instances of each product in an image, and return bounding boxes with counts per class. This is the only option that supports custom, multi-class object detection and counting from user-provided training data.
What should I do if I get this AI-900 question wrong?
Review custom object detection trains models to identify specific objects in images., then practise related AI-900 questions on the same topic to reinforce the concept.
What is the key concept behind this question?
Custom object detection trains models to identify specific objects in images.
About these practice questions
Courseiva creates original exam-style practice questions with explanations and wrong-answer analysis. It does not publish real exam questions, exam dumps, or protected exam content. Learn why practice questions differ from exam dumps →
Share a tip, memory trick, or ask about the reasoning behind this question. Do not post real exam questions, leaked content, braindumps, or copyrighted exam material. Comments are moderated and may be removed without notice.
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.
Question Discussion
Share a tip, memory trick, or ask about the reasoning behind this question. Do not post real exam questions, leaked content, braindumps, or copyrighted exam material. Comments are moderated and may be removed without notice.
Sign in to join the discussion.