Field Level / Zonal Recognition (OCR,ICR)

Language:
EN
Product-Line:
FineReader Engine, Cloud OCR SDK
Version:
9.x, 10, 11
Type:
Scenarios/Tasks
Category:
Recognition, Verification & User Interaction
  • Zonal OCR or Field Level recognition is a special recognition scenario where the text recognition quality is very important. The information that is extracted is used in further business processes. Typically the information that is OCRed is only a small portion of the scanned document, e.g. the basic lines on a banking form.
  • The “region of interest” of the relevant text is well known in advance; often there are only a few different document variants.
  • In pure “document conversion” or “fulltext indexing” scenarios the automatic layout analysis is very important, because the text areas have to be localized first. In Field Level OCR scenarios the zones are defined via code/API.
  • ABBYY SDKs offer additional features that are important for a good field level recognition implementation:
    • To get the best text result, the image quality of the snippet has to be as good as possible. The following routines might be applied:
      • Auto-detection of page orientation
      • Image despeckling
      • Lines straightening
      • Removing motion blur and ISO noise from digital photos
    • OCR Voting API to provide the developer all details about the OCR result of a text
    • FineReader Engine 10 introduced a Processing Profile FieldLevelRecognition - so there is pre-testet set of OCR engine settings, tuned for Zonal OCR

Zonal Recognition in ABBYY Products

  • For field level recognition normally OCR/ICR SDKs are used, so FineReader Engine is the SDK where this topic belongs to.