Bitzu

AI-901 ยท Information extraction with Content Understanding

How does Content Understanding extract information from images?

Describing images and extracting defined fields such as a summary or specific attributes from pictures, and how a field schema guides the result.

Key points

  • You define a field schema (field names, types and descriptions) and the analyzer fills it from each picture, for example a summary, a defect description, or the count of products on a shelf.
  • The prebuilt prebuilt-imageSearch analyzer returns a one-paragraph description for search and RAG.
  • Image analyzers are not tuned for text-heavy pictures: for mostly text, use a document schema.

Source: https://learn.microsoft.com/en-us/azure/ai-services/content-understanding/image/overview

Practice questions

Single choice

Which prebuilt analyzer returns a one-paragraph description of an image for search and RAG?

  1. prebuilt-read
  2. prebuilt-imageSearch
  3. prebuilt-audioSearch
  4. prebuilt-callCenter
Show answer

Answer: B. prebuilt-imageSearch

prebuilt-imageSearch generates the image description. prebuilt-audioSearch and prebuilt-videoSearch handle other modalities, and prebuilt-read only does OCR on documents.

Select all that apply

What benefits does defining a field schema for images bring?

  1. Output is structured data that integrates with databases and business systems
  2. Only the fields you need are produced, reducing irrelevant data
  3. Targeted fields help focus accuracy on the most important data points
  4. The schema retrains the underlying model on your photos
  5. Every extracted value is guaranteed correct
Show answer

Answer: A. Output is structured data that integrates with databases and business systems, B. Only the fields you need are produced, reducing irrelevant data, C. Targeted fields help focus accuracy on the most important data points

A schema returns only the needed fields in a structured form, which is easy to feed into databases or CRM/ERP, and focuses the model on what matters. It does not retrain a model or guarantee perfect results.

True or false

True or false: an image analyzer is the best tool when your main goal is to read and analyze the text on scanned pages.

Show answer

Answer: False

False. Image analyzers are not optimized for text-centered analysis; use a document field extraction schema for that.

Practice 8 questions on this topic in Bitzu

Short lessons, instant explanations and spaced review of what you get wrong. Free on iOS and Android.

Download on the App StoreGet it on Google Play

More AI-901 topics in this module