AI-103 Practice Question 98: Image Tags Photo Search
Microsoft Azure AI Apps and Agents Developer Associate (AI-103) · Test 10 · question 8 of 10
A stock-photo marketplace wants to auto-generate searchable keywords for every photo it hosts, dozens of loosely related single words per image, such as "beach," "outdoor," "umbrella," "sand," and "vacation," so a text search for any one of those words surfaces the photo. There's no labelled training data, the marketplace's subject matter changes constantly, and the team needs a flat list of relevant words for a search index, not full sentences.
Which Image Analysis 4.0 visual feature fits this requirement?
Show answer and explanation
Answer:
Content tags are defined for exactly this output: "Image Analysis can return content tags for thousands of recognizable objects, living beings, scenery, and actions... Tagging is not limited to the main subject... but also includes the setting... furniture, tools, plants, animals, accessories, gadgets, and so on." That's a flat list of loosely related keywords, precisely what a search index needs, generated with no labelled training data and no fixed category list to maintain as inventory changes.
- B is wrong because dense captions return full sentences describing regions of the image, which is more verbose than a keyword index needs and doesn't produce the flat tag list the scenario asks for. - C is wrong because a Custom Vision classifier requires labelled training examples for a fixed set of categories the team would have to define and keep current, the opposite of "no labelled training data" for an ever-changing photo library. - D is wrong because OCR reads text physically printed in the image; it cannot identify a beach, an umbrella, or sand unless those words happen to appear as text in the photo.
Reference: Content tags - Azure Vision in Foundry Tools