When creating a new layout, you will be asked to choose which layout type you would like to create. Hyperscience supports layouts for three different types of documents: Structured, Semi-structured, and Additional.
You will need to create a layout for the document that you would like to extract information from. For more information on distinguishing different document types, see Understanding Document Types.
Structured layouts
Layouts for Structured documents are used to identify and capture data from pages with a predictable structure. If you have a blank form, then you need to create a layout for a Structured document. All of the pages that match a Structured layout contain information you want to extract in the same location with the same labels.
For example, a stellar Structured layout candidate is a federal government form. Since it is a standardized form, it will have the same data fields, reliably located in the same places on the page.
You can create layout variations for Structured layouts, which allows you to group very similar layouts together with a shared field list and field customizations. For example, if you have a form that changes every year or that may vary slightly based on the software vendor who created it, you can create a layout variation for each instance of the form and group them together as a layout in Hyperscience (e.g., a layout named “HCFA” with variations named “HCFA 2014” and “HCFA 2015”). To learn more, see Creating Structured Layouts and Adding a Variation to a Layout.
If you have highly variable Structured forms of the same type, and you cannot create a layout variation for each individual form, we recommend building Semi-structured layouts as an alternative. For example, if a document you’re processing is slightly different from all of your Structured layout variations, you can classify this document as a Semi-structured form. Thus, you will be able to increase automation in terms of Field ID and transcription.
Semi-structured layouts
Layouts for Semi-structured documents are used to identify and capture data from pages without a consistent structure or page geography. While the information that you want to extract from these pages is consistent (e.g., identification number, address, etc.) the location of this information may vary and can have different names or labels.
For example, a superb Semi-structured layout is a paystub or invoice. There are certain pieces of information that are expected to appear on the page, but the locations may differ widely across variations of the page.
More information about Semi-structured layouts can be found in Creating Semi-structured Layouts.
Additional layouts
Layouts for documents that contain no fields for extraction, just categorization, are called Additional Layouts. These layouts are helpful for pages within a submission that should be included in the submission output for context.
For example, a cover sheet would be a good Additional layout. This may not have fields, but it does provide context about the submission itself.
To learn more, see Creating Additional Layouts.