AI-900 Practice Question: Describe features of computer vision workloads on Azure
What is the Azure AI Vision service's 'Image Analysis 4.0' major new capability compared to previous versions?
⚠ Common exam trap
Candidates often confuse Image Analysis 4.0's new Florence model with general AI improvements, mistakenly thinking video support or larger file sizes are the headline feature, when the core innovation is the foundational model's advanced image understanding.
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
✓
The Florence foundation model enabling detailed captions, dense captioning, background removal, and multimodal embeddings
Image Analysis 4.0 introduces the Florence foundation model, which significantly enhances image understanding capabilities. This model enables detailed captions, dense captioning (generating captions for multiple regions within an image), background removal, and multimodal embeddings that align images and text in a shared vector space. These features go far beyond the classification, object detection, and OCR capabilities of version 3.x.
Answer analysis
Option-by-option breakdown
For each option: why learners choose it and why it is or isn't the right answer here.
- ✗
Support for processing video files, which was not available in version 3.x
Why it's wrong here
Video processing is outside the scope of the Image Analysis API, which accepts still images as input; separate services like Azure Video Indexer or Azure Video Analyzer handle video content. The claim that video support was entirely absent in 3.x is misleading, and even if video were a consideration, Image Analysis 4.0 does not add video ingestion. The real new capability in 4.0 is the Florence foundation model, which enriches still-image understanding with detailed captions, dense captioning, background removal, and multimodal embeddings.
- ✓
The Florence foundation model enabling detailed captions, dense captioning, background removal, and multimodal embeddings
Why this is correct
Image Analysis 4.0 is powered by the Florence foundation model, a large-scale vision model that delivers the signature capabilities of this API version. These include detailed one-sentence captions, dense captioning that describes multiple objects and regions within a single image, background removal via foreground segmentation, and multimodal embeddings that map images and text into a shared vector space. This is the major advancement over version 3.x, which relied on narrower, task-specific models and could not provide the same depth of semantic understanding.
- ✗
Support for the first time for color analysis features in images
Why it's wrong here
Color analysis, such as extracting dominant colors, accent colors, and black-and-white detection, was already present in Computer Vision 3.x image analysis features. It is not a newly introduced capability in Image Analysis 4.0, and 4.0's core novelty is instead the Florence foundation model, which powers dense captioning, region-based descriptions, and background removal. Therefore, saying 4.0 adds color analysis for the first time mistakes a legacy feature for the version's actual foundation-model-driven advancement.
- ✗
The ability to process images larger than 4MB for the first time
Why it's wrong here
Although Image Analysis 4.0 does support somewhat larger file sizes than the older 3.x limit (commonly cited as 6 MB versus 4 MB), this is an incremental technical specification, not the architectural breakthrough. The phrase “for the first time” is imprecise, since file-size limits can be adjusted without changing the underlying model. The defining leap in 4.0 is the integration of the Florence foundation model, enabling semantic capabilities such as open-ended image descriptions and multimodal embeddings rather than merely accepting bigger images.
Go deeper
Related to this question
Learn chapter
Regression and Classification
Key term
Model
In IT and AI, a model is a trained mathematical representation that learns patterns from data to make predictions or decisions.
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.