AI-103 Practice Question 18: Dense Captions Alt Text
Microsoft Azure AI Apps and Agents Developer Associate (AI-103) · Test 2 · question 8 of 10
A news agency is retrofitting alt text onto a photo archive. For each photograph the accessibility team wants a single sentence describing the picture as a whole, plus separate one-sentence descriptions of the individual regions within it, each with coordinates so a human editor can confirm the description matches the right part of the frame.
Which Image Analysis 4.0 visual feature should the job request?
Show answer and explanation
Answer:
Dense Captions generates one-sentence descriptions of up to 10 different regions of the image *in addition to* describing the whole image, and returns bounding box coordinates for each described region, both halves of the requirement in a single call. Two operational notes matter for an archive job: image captioning is English only and is available only in certain regions, so the Azure Vision in Foundry Tools resource has to be created in one of them.
- B is wrong because Caption returns exactly one sentence for the entire image and no coordinates, so editors get nothing to verify region by region. - C is wrong because Tags returns individual content words, not sentences, and is not suitable as alt text on its own. - D is wrong because object detection returns bounding boxes with short class labels such as "chair"; the coordinates are there, but the descriptive sentence the alt text needs is not.
Reference: Image captions - Image Analysis 4.0