Reading Text in Images and Documents

What are some examples of how Optical Character Recognition could be used in real world applications?

• Extract text from bills and invoices

• Transcribe text from one script to another (e.g. Japanese script to Latin script)

• Extract text from product

• Extract words from audio files

What API does Computer Vision offer for reading text?

• The Vision API

• The Read API

• The Language Reader API

• The Text API

What is one reason why an intake desk might want to use the Form Recognizer service instead of manually re-entering data from intake form

• Form Recognizer services provide the building blocks for automation by using intelligent services to extract data at scale.

• Form Recognizer services will not make mistakes that are often made by humans when re-entering data manually

• Form Recognizer services can replace people at intake desks, reducing the cost of running a business.

• Form Recognizer should only be used on typed forms, and so wouldn’t be useful at an intake desk.

Form Recognizer can be used on what formats?

• JPG or PNG

• PDF

• TIFF

• DOCX

What is an example of how the Business Cards prebuilt Form Recognizer model could be used?

• Processing paperwork for each card owner in real time.

• Automatically creating contacts from images to integrate data with a CRM

• Analyzing consumer behavior with business card drop offs

• Improving business card design through comparisons of collected cards

In an attempt to store the handwritten notes of an English department at a University, you’re tasked with creating an application that will allow researchers to take a picture of the documents and store the text digitally. Which API would you use and why?

Answer explanation

Your Answer: Should include the Read API because it can read handwritten text in English

Describe one detail besides the actual text in the JSON response of the Read APIs.

Answer explanation

Your Answer: Should include one of the following: page, angle, bounding box, size, appearance style, confidence score

What are the three services that Form Recognizer provides that improve the service over a simple Read API?

Answer explanation

Your Answer: Should include Layout Service that returns information in bounding boxes, text content, tables, selection marks, and document structure, Prebuilt Models that currently support receipts, business cards, and invoices, and Custom Models that include unsupervised and supervised learning for training custom models for use in unlabeled and labeled forms.

Describe an example of how the prebuilt Receipt model of the Form Recognizer service could be used.

Answer explanation

Your Answer: Could include streamlining business expense reporting processes, automating auditing and accounting tasks, and analyzing consumer behavior and shopping trends.

You are creating a receipt reader application for a customer and decide to use the prebuild Receipt model. What happens when you call the Analyze REST API function?

Answer explanation

Your Answer: Should include a description of how the result ID is returned and then used to call the Get Analyze Result function to actually retrieve the results. The results will be in one JSON response with two nodes: readResults and documentResults, which contain all of the recognized text and the form-specific values.