OCR skill JSON schema

The schema described in this article determines the structure of JSON files containing document processing results for an OCR skill. These files store information about graphic elements, text elements, tables, and lists detected on the image, as well as information about the estimated document structure: all detected text is split into paragraphs, each of which is assigned a specific function. Paragraphs are grouped into logic sections.

All object properties, their data types, and their allowed values are listed in the JSON schema, hence we recommend familiarizing yourself with it in order to make it easier to set up tools to analyze JSON files. You can download the JSON schema for OCR skills here.

The root object describes the whole recognized document. General information about the document is stored in the layout and content objects. The layout object describes image parameters and all detected graphic and text elements, while the content object describes the data components of the document and their place in its logic structure.

The table below lists all root object properties:

Property Data type Description
version* string

The version of the JSON schema document.

Set to the following by default: Vantage OCR.Skill JSON output v1.0

producer* string

The source of the JSON file.

Set to the following by default: ABBYY Vantage OCR.Skill

languages string array A list of all languages detected in the document.
layout object Physical (layout) structure of the document.
content object Content (logic structure of the document).

* Indicates a required property.

22.12.2023 12:36:42

Usage of Cookies. In order to optimize the website functionality and improve your online experience ABBYY uses cookies. You agree to the usage of cookies when you continue using this site. Further details can be found in our Privacy Notice.