Boot slides xxl
- 1. Some Customers OEM Direct Customers o 1
- 2. Value Proposition Collect documents from … classify … and feed them into various sources …. them, extract data business applications … Transform Pixels into Content Server Actionable Information TCP, IDM, VRD, C 360, P360, … SharePoint Automatically with OCR, ICR, IDR
- 3. Use Cases Enable Digital Workflow Qualified Electronic Digital Mail Room Document Scanning Documents Make into Electronic Files Money Document Backfile Conversion and Data Capture Transaction and Process Save Management Money Service Centers Reduce Manual Ad-Hoc capturing Keying Enable Process Auto- mation Slide 3 Copyright © Open Text Corporation. All rights reserved.
- 4. Overview Components of OCC Open Text Capture Center Monitoring Configuration Enterprise Scan Business Client Application IM Dispatch EX Fax, Email, FTP site, Archive Network Folder SharePoint Recognition Validation Copyright © Open Text Corporation. All rights reserved.
- 5. Recognition: Different Document Types Structured Semi-Structrured Unstructured Documents Documents Documents > Forms > B2B Correspondence > C2B Correspondence Data at fixed Data is in logical Data may be positions groups but positions anywhere in the are unknown. document. Copyright © Open Text Corporation. All rights reserved.
- 6. Steps in Document Recognition Separation: Splitting a batch of images into individual (multi-page) documents … Classification: Identification of document type „Invoice” „Delivery Note“ „Order“ „Others“ Extraction: Searching for basic information on the document Date: Jan 21. 2012 Amount: 332,29 $ Order-Nr: X-44277 Invoice Supplier: Mueller & Friends Interpretation: Enhancing of extracted data with context information Date: Jan 21. 2012 21012012 Amount: 332,29 $ 332.29 USD Order-Nr: X-44277 X-44277 Supplier: Mueller & Friends K441258-3 Slide 6 Copyright © Open Text Corporation. All rights reserved.
- 7. Free Forms Extraction OCR Analyze Extract Normalize Verify “10/24/10” 24.10.2010 Valid Format (US) period 3 Date Order_Date = “10/24/10” Day = 24 Month = Oct Year = 2010 yes ? no CLERK: 12 DATE SHIPPED ORDER DATE Word 10/30/02 10/24/02 Manual 24.10.2010 Export COVER CODE keying Number PA Turn pixels into Find meaningful entities Find the correct date Decompose string into Check against business characters w/ optical and tag them among all detected subunits and re-format as and plausibility rules character recognition alternatives required Copyright © Open Text Corporation. All rights reserved.
- 8. Long Term Effect of Free Forms Techniques OCC Approach Form Adaptive Combining (template) Free Form (=learning) Free Forms based Recognition technology and adaptive Recognition Increase of recognition rate during production time Copyright © Open Text Corporation. All rights reserved.
- 9. Invoice Data Extraction Supplier Invoice Number Invoice Date Order number Delivery note Line items Net Amount Currency Total amount Copyright © Open Text Corporation. All rights reserved.
- 10. Supported Countries in Knowledge Base United Belgium Australia Norway Czech Russia States Republic Austria France Denmark Singapore Slovakia* New Zealand Netherlands Finland Spain Poland Slovenia United Hungary Sweden Canada Romania* Kingdom Germany Italy Switzerland Portugal Bulgaria* * Header data only
- 11. Purchase Order Processing with OCC Rendition Web form Sever Capture Center Enterprise Reco- Validation Case360 Scan gnition Master Data Knowledge Base Content Tempo Sever 11
- 12. OCC – Your Benefit Reduce Operating Costs $ Automating manual tasks Deploying a single input management platform Reducing paper filing/storage Improve Information Quality Q Improving information sharing Leveraging a common set of business rules Reducing errors Accelerate Business Processes Shortening cycle times Reducing exception processing Enhancing customer relationships Improving knowledge worker productivity §§ Compliance Ensuring compliance / auditability Improving visibility into business processes Improving litigation preparedness Copyright © Open Text Corporation. All rights reserved.