Classify, separate, sort and extract
Parashift provide support for Ephesoft’s open source Linux Community Edition, for intelligent document capture and mailroom automation.Presenting the next generation of optical character recognition technology (OCR), featuring unique Intelligent Character Recognition (ICR), Ephesoft is driven with the mission to help businesses operate more efficiently by providing them with the automated capability of document capture and management. In addition to scanning forms, Ephesoft has the built-in intelligence to sort and classify the documents (such as invoices, contracts, proposals and letters). The system then extracts relevant information from these documents to execute workflows mirroring your business processes. When integrated with other tools such as ProcessMaker, Ephesoft can activate a set of automated internal processes, such as reviews and approvals. Ephesoft Features:
- Open Source
- Unlimited users and copies
- On premise and cloud deployment
- Automated classification with human exception based validation
Ephesoft Technical Specifications
Ephesoft is a Document Capture engine which has a pluggable infrastructure.Why would I need it?
You would need ephesoft if:
- You have any forms that are hand written by your customers and need a fast and effective way to ingest these
- You want to capture paper records or migrate to a digital by default workplace
- You want to automatically classify documents and use them within workflows
How does it work?
Ephesoft is built using open source technologies. Ephesoft provides a process for capturing incoming documents such as paper, fax, email attachments.
- Data entry operators review and validate the scanned/imported documents. These users mainly use keyboard, mouse or keyboard shortcuts to have high efficiency.
- Supervisors do system level operations like reporting, setup, configuration, etc.
- Administrators configure the batch class and How Ephesoft interacts with Document Management Systems, Business Process Management Systems or Databases.
Ephesoft is built on top of a workflow engine. Each step in the workflow is called Plugin and independently responsible for one specific operation. A Plugin might be responsible of OCRing the page and another plugin might be responsible for exporting documents to a repository. Plugins are grouped into sub workflow containers which are called Modules. For example all plugin that are used to extract meta data from documents such as Free Form extraction or Zonal OCR/ICR, or table/Line item extraction plugins can be found in a module called Extraction.Overall System Architecture
The below diagram shows the Overall Workflow and how Plugins and Modules are implemented. Workflow can be followed from left to right. Documents are imported by the plugins in the Import Module, each page is analysed by various plugins in Page Processing Module, Document boundaries are identified by the plugins in the Document Assembler Module, Operators review the documents classification results (if necessary) in Document Review, Meta Data in extracted from Documents based on document type using the plugins in Extraction Module, Operators validate the extracted values in Document Validation and finally Documents are exported to their destinations. Ephesoft Document Capture and Mailroom Automation workflow consists of automatic modules and manual modules.
The automatic modules, shown in green, does not require any user interaction to execute plugins. Ephesoft Server automatically processes the documents as soon as they are available.
The manual modules, shown in blue, is used only when needed. If the document classification is not needed Ephesoft automatically skips this step in the workflow. Same is also valid for Document Validation. Both Document Review and Validation are designed for a very specific purposes and they provide unique features to help their functions. Review is designed for reviewing the classification results/exceptions. If Ephesoft is not sure about a document, it will ask users to correct or verify the document type in this User Interface. Validation is designed for reviewing extracted meta data fields. If the fields/data is not confidently captured or has a missing information, it will ask users to complete or verify the captured data. Both Review and Validation also supports automatic and manual batch selection. In automatic batch selection, the system automatically selects the next available batch for processing. The user, based on his rights, can either review the batch or can both review and validate it. In manual batch selection, the user is shown a list of all the batches available for review and validation in the grid and he/she can select the batch (to work).