Optical Character Recognition engines: UiPath Screen OCR, Tesseract, Omnipage, scale factors, and image preprocessing.
Traditional / Ad-Hoc Approach
Manual human clicks, brittle unmanaged scripts, zero centralized control, and missing fault tolerance.
Enterprise UiPath Standard
Production-grade execution, dynamic Unified Target selectors, automated retry recovery, and centralized governance.
Optical Character Recognition (OCR) activities analyze raw pixel patterns on screen or inside document images to reconstruct characters into editable digital text. Key Sub-topics Breakdown: • Subtopic 1: OCR Engines: UiPath Screen OCR, Tesseract OCR, Omnipage OCR, Google Cloud, and Microsoft Azure OCR • Subtopic 2: Surface Automation: Extracting text from flat rasterized pixels, remote Citrix streams, and scanned PDFs • Subtopic 3: Image Preprocessing: Scale factors, binarization, contrast inversion, and language profiles In enterprise automation environments, OCR Activities guarantees reliable execution within the UiPath ecosystem. Key Sub-topics Breakdown: 1. OCR Engines: UiPath Screen OCR, Tesseract OCR, Omnipage OCR, Google Cloud, and Microsoft Azure OCR 2. Surface Automation: Extracting text from flat rasterized pixels, remote Citrix streams, and scanned PDFs 3. Image Preprocessing: Scale factors, binarization, contrast inversion, and language profiles
When automating remote Citrix/RDP desktops or scanned physical documents, underlying UI controls do not exist; OCR is the only way to read text.
The activity takes a screen clipping, applies image preprocessing (grayscale, contrast threshold), passes pixels to the OCR engine, and returns recognized text.
Zero unmanaged credentials, explicit timeout ceilings, and structured audit logging.
This enterprise architecture diagram illustrates the execution lifecycle and component boundaries for OCR Activities:
ENTERPRISE AUTOMATION PIPELINE (OCR Activities):
┌────────────────────────────────────────────────────────────────────────┐
│ Design & Governance Plane: UiPath Studio & Orchestrator │
│ • Packages (.nupkg) • Modern Folders • Credential Assets / Queues │
└───────────────────────────────────┬────────────────────────────────────┘
▼
┌────────────────────────────────────────────────────────────────────────┐
│ Execution Plane: UiPath Robot (Attended / Unattended Agents) │
│ ┌───────────────────────────────────┐ ┌────────────────────────────┐ │
│ │ UI Automation (Simulate / Chromium)│ │ Data & API Processing │ │
│ │ Unified Target & Object Repo │ │ LINQ / HTTP Web Requests │ │
│ └───────────────────────────────────┘ └────────────────────────────┘ │
└───────────────────────────────────┬────────────────────────────────────┘
▼
┌────────────────────────────────────────────────────────────────────────┐
│ Enterprise Exception Handling & Telemetry Sinks │
│ • REFramework States • Try-Catch / Global Handler • Audit Logs │
└────────────────────────────────────────────────────────────────────────┘At a global enterprise handling over 75,000 monthly transactions, operational teams implemented OCR Activities to resolve critical production bottlenecks:
Manual intervention caused processing delays and human error in mission-critical transactions.
Deployed OCR Activities with automated retry rules, dynamic error recovery, and end-to-end audit logging.
Eliminated 98% of manual touchpoints, achieved sub-second execution latency, and maintained 99.9% uptime.
Evaluate the trade-offs, performance SLAs, and production constraints when deploying OCR Activities: