Machine Learning

Automating Document Processing with Amazon Textract - From OCR to Form and Table Extraction

Go beyond OCR with structural recognition of forms and tables to automatically extract data from invoices, receipts, and identity documents. Also covers integrating human review with A2I.

約 4 分で読めます

Textract's API Suite

Textract is a machine learning-based document analysis service that provides structured data extraction beyond traditional OCR. DetectDocumentText is the basic OCR feature that extracts text from images and PDFs at the line and word level. AnalyzeDocument is an advanced analysis feature that recognizes forms (key-value pairs) and tables (row-column structures). For example, from an application form containing "Name: Taro Yamada," it automatically pairs the key "Name" with the value "Taro Yamada." AnalyzeExpense is an API specialized for invoices and receipts, extracting vendor names, invoice dates, total amounts, tax amounts, and line items as structured data. AnalyzeID extracts information such as name, date of birth, and address from driver's licenses and passports.

Asynchronous Processing for Large Document Volumes

The synchronous API handles single-page images, but for multi-page PDFs or large document volumes, you use the asynchronous API. Start processing with StartDocumentTextDetection or StartDocumentAnalysis, and a completion notification is sent to an SNS topic. The standard pattern is an event-driven architecture where a Lambda function receives the notification and retrieves results with GetDocumentTextDetection or GetDocumentAnalysis. By building a pipeline that triggers Lambda on S3 document uploads to call Textract and store extraction results in DynamoDB, you can fully automate document processing.

Improving Accuracy and Human Review

Textract assigns a confidence score (0-100%) to each extracted field. For results where confidence falls below a threshold, you can use Amazon Augmented AI (A2I) to route them to a human review workflow. Reviewers use the A2I console to compare the original document side by side with the extraction results and make corrections. Corrections are accumulated as feedback and used to improve the quality of subsequent processing. Textract's Queries feature lets you specify natural language questions (e.g., "What is the patient's name?") to extract specific information from documents, handling documents with irregular form structures. For a comprehensive study of Textract's algorithms, check out technical books on Amazon.

Textract Pricing

Textract uses per-API pay-as-you-go pricing. DetectDocumentText (OCR) costs approximately $0.0015 per page, AnalyzeDocument (forms and tables) approximately $0.015, AnalyzeExpense (invoices) approximately $0.01, and AnalyzeID (identity documents) approximately $0.01 per page. The Queries feature adds approximately $0.005 per query on top of the approximately $0.015 per page. For large document volumes, a two-stage approach of first running DetectDocumentText for OCR and then applying AnalyzeDocument only to pages requiring structured data extraction optimizes costs.

Summary

Textract goes beyond traditional OCR by understanding document structure to extract data. It provides specialized APIs for different document types including forms, tables, invoices, and identity documents, significantly reducing manual data entry work. Integration with A2I enables incorporating human review, making it suitable for business processes that demand high accuracy.

Related Services

Amazon TextractAn OCR service that automatically extracts text, tables, and form data from documentsAmazon S3A scalable object storage serviceAWS LambdaA serverless compute service that runs code without server management

Related Articles

Implementing Image and Video Analysis with Amazon Rekognition - From Label Detection to Custom ModelsLearn how to implement label detection, facial analysis, and text detection using pre-trained APIs, and build domain-specific image recognition models with Custom Labels.Implementing Natural Language Processing with Amazon Comprehend - Sentiment Analysis and Entity ExtractionLearn about sentiment analysis, entity extraction, and building custom classification models with Comprehend.Implementing Speech-to-Text with Amazon Transcribe - Real-Time Conversion and Custom VocabulariesProvides both batch and real-time speech-to-text transcription, with custom vocabularies to improve accuracy for industry-specific terms. Also covers quality management for contact centers with Call Analytics.

More on This Topic

GPU-Based Machine Learning Training with AWS Batch - Cost-Efficient Large-Scale TrainingRun GPU training with your existing Docker containers, and cut costs by up to 90% using Spot Instances and checkpointing. Includes guidance on when to choose Batch over SageMaker.Using Claude on Amazon Bedrock - Model Selection, Prompt Design, and Cost OptimizationCompares the Anthropic Claude models available on Amazon Bedrock, provides model selection guidelines by use case, and covers prompt design best practices and cost optimization.Building RAG Applications with Amazon Bedrock Knowledge Bases - Implementing Retrieval-Augmented GenerationAutomatically index documents on S3 and unify search and generation with the RetrieveAndGenerate API. Covers chunking strategy selection and safety enforcement with Guardrails.Getting Started with Quantum Computing on Amazon Braket - Designing and Simulating Quantum CircuitsPrototype for free with local simulators, then run quantum circuits on IonQ and Rigetti hardware. Covers implementing VQE and QAOA with hybrid jobs.Privacy-Preserving ML with AWS Clean Rooms ML - Build Models Without Sharing DataLearn how to build lookalike models with Clean Rooms ML, apply differential privacy, and leverage the results for ad targeting.Implementing Natural Language Processing with Amazon Comprehend - Sentiment Analysis and Entity ExtractionLearn about sentiment analysis, entity extraction, and building custom classification models with Comprehend.Building Conversational Bots - Natural Conversation Interfaces with Amazon Lex and PollyLearn how to build conversational bots using Amazon Lex and Amazon Polly.Demand Forecasting - Predicting the Future from Time Series Data with Amazon ForecastInput historical time series data and related variables to automatically build ML-based demand forecasting models. This guide covers forecast accuracy evaluation metrics and patterns for leveraging forecast results through S3 and QuickSight integration.

Similar Articles and Services

Document Text Extraction - Intelligent Document Processing with Amazon TextractLearn how to automatically extract text, tables, and form data from documents with Amazon Textract, and build natural language processing pipelines by integrating with Amazon Comprehend. This article covers automation patterns for invoice processing and contract analysis.Amazon TextractA machine learning-based document analysis service that automatically extracts text, handwriting, tables, and form key-value pairs from documentsAutomatic Data Extraction from Documents with Amazon Textract - OCR, Table Analysis, and Form RecognitionLearn how to extract text from documents, analyze table structures, and extract key-value pairs from forms using Textract.Amazon KendraA machine learning-powered enterprise search service that returns accurate answers to natural language questions