easyMultiple Choice
AI-102 Practice Question: A company uses Azure Custom Vision to classify…
A company uses Azure Custom Vision to classify images of defective parts. After deploying the model, the accuracy is low. The team only has 10 images per class. What is the most effective way to improve accuracy?
⚠ Common exam trap
Watch out — candidates often assume increasing epochs or changing the algorithm will fix low accuracy, but the real bottleneck is insufficient and non-diverse training data, which is the most common cause of poor Custom Vision model performance.
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
✓
Add at least 50 more images per class with variations.
Azure Custom Vision relies on deep learning models that require a sufficient number of diverse training images to generalize well. With only 10 images per class, the model is severely underfit and prone to overfitting; adding at least 50 more images per class with variations in lighting, angle, and background provides the necessary data diversity to improve accuracy significantly.
Answer analysis
Option-by-option breakdown
For each option: why learners choose it and why it is or isn't the right answer here.
- ✗
Use a different classification algorithm.
Why it's wrong here
Custom Vision selects its classifier architecture automatically; swapping algorithms is not a configurable lever and cannot compensate for ten images per class. Algorithm choice matters when datasets are adequate but the domain demands a specific model type, which is not the constraint here.
- ✓
Add at least 50 more images per class with variations.
Why this is correct
Ten images per class is far below what Custom Vision needs to generalise; adding at least 50 varied images per class gives the model enough examples to learn distinguishing features, directly addressing the small-dataset constraint that is causing the low accuracy.
- ✗
Reduce the image resolution to speed up training.
Why it's wrong here
Lowering resolution discards the fine defect detail the classifier needs, worsening accuracy rather than improving it. Resolution reduction is chosen to cut training time on large datasets, but the actual constraint is ten images per class, which more data addresses.
- ✗
Increase the number of training iterations (epochs).
Why it's wrong here
Extra epochs cannot manufacture information absent from ten images per class; the model simply overfits the same samples. Increasing iterations suits underfit models trained on sufficient data, whereas the stem's constraint is dataset size, requiring additional labelled images.
Go deeper
Related to this question
About these practice questions
Courseiva writes every AI-102 question from scratch — 761 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-102 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-102 exam.