Computer Vision Services

Question 1

A construction company needs to automatically verify if workers are wearing helmets and high-visibility vests in construction zones to comply with safety regulations. Which solution requires the LEAST development effort?

Accepted Answer

Amazon Rekognition DetectProtectiveEquipment

Answer

Amazon Rekognition Custom Labels

Answer

Amazon SageMaker Object Detection

Answer

AWS Panorama with a custom model

Question 2

A hospital wants to digitize old handwritten medical forms. They need to specifically extract 'Diagnosis' and 'Admission Date' from documents with varying layouts. Which Textract feature is the most efficient?

Accepted Answer

Amazon Textract Queries

Answer

Amazon Textract Forms (Key-Value Pairs)

Answer

Amazon Textract Tables

Answer

Amazon Comprehend Medical

Question 3

A factory needs to inspect parts on an assembly line in real-time. Internet connection is unstable, and latency must be under 20ms to stop the belt if defects are found. Which architecture is most suitable?

Accepted Answer

Deploy a Computer Vision model on AWS Panorama Appliance

Answer

Send video to Amazon Rekognition Video via Kinesis Video Streams

Answer

Use Amazon Rekognition Custom Labels with cloud processing

Answer

Upload images to S3 and trigger a Lambda function with Rekognition

Question 4

An auto parts store wants to identify specific engine parts (Brand X spark plugs, Brand Y filters) using photos. The generic Rekognition model only identifies them as 'Car Part'. What should they do?

Accepted Answer

Train a model with Amazon Rekognition Custom Labels

Answer

Use Amazon Rekognition Image with lower confidence labels

Answer

Implement Amazon Textract to read labels on the parts

Answer

Use Amazon SageMaker JumpStart for a foundation model

Question 5

A financial app needs to process thousands of scanned restaurant receipts and extract 'Total', 'Date', and 'Vendor Name' for expense reports. Receipt formats vary widely.

Accepted Answer

Use Amazon Textract AnalyzeExpense

Answer

Use Amazon Textract AnalyzeDocument with FeatureType TABLES

Answer

Use Amazon Rekognition Text in Image

Answer

Use Amazon Comprehend to extract entities

Question 6

You are designing a surveillance system for an airport. You need to track the path of a person of interest across multiple cameras in real-time. Which services should you combine?

Accepted Answer

Amazon Kinesis Video Streams and Amazon Rekognition Video

Answer

AWS Lambda and Amazon Rekognition Image

Answer

Amazon S3 and Amazon Rekognition Video Stored Video Analysis

Answer

AWS Panorama and Amazon SageMaker

Question 7

A content management system (CMS) wants to automatically generate descriptive alt-text for millions of user-uploaded images to improve accessibility.

Accepted Answer

Amazon Rekognition DetectLabels

Answer

Amazon Rekognition DetectText

Answer

Amazon Textract AnalyzeDocument

Answer

Amazon Transcribe

Question 8

You need to extract data from passports for a 'Know Your Customer' (KYC) process. Which service offers the highest accuracy and validates the document structure?

Accepted Answer

Amazon Textract AnalyzeID

Answer

Amazon Textract DetectDocumentText

Answer

Amazon Rekognition Face Comparison

Answer

Amazon Comprehend PII Reduction

Question 9

Your system uses Rekognition to detect inappropriate content. Occasionally, the system flags false positives on artistic images. You want to send these low-confidence images to a human team for review. Which service should you integrate?

Accepted Answer

Amazon Augmented AI (A2I)

Answer

Amazon Macie

Answer

Amazon CloudWatch Alarms

Answer

AWS Systems Manager OpsCenter

Question 10

A financial firm wants to analyze complex tables in annual reports (PDFs) where rows span multiple lines and there are merged cells. DetectDocumentText is not maintaining the structure correctly.

Accepted Answer

Use AnalyzeDocument with FeatureType TABLES

Answer

Use AnalyzeDocument with FeatureType FORMS

Answer

Convert PDF to image and use Rekognition

Answer

Use Amazon Kendra to index the document