AutoRec Uno  |   AutoRec OMR  |   AutoRec Forms  |   AutoRec WEB  |   AutoRec CPS  |   AutoRec IRIS
Multiple User  -  AUTOREC WEB
Product Overview

For high volume data capture AutoRec Web can be installed on a wide area network. Multiple scanning stations and proofreading stations can be located at diverse geographical locations, while the recognition and designing happens on a centralized server. AutoRec Web is designed for web level solutions, processing from several thousand to hundreds of thousands of documents each day.

AutoRec Web is a multifaceted data capturing software. It incorporates five recognition technologies. In addition to handwriting recognition, It can recognize machine print, ticks, crosses, marks and barcodes. AutoRec Web also is capable of cropping photographs and signatures from a document.

Methodology

Scanned images get uploaded onto a central server. Central server reads the images and sends the recognized text back to a proofreading station which might be located anywhere in the world. After proofreading the corrected data gets sent to a central server and can be made available in any format.The data is encrypted and compressed through hardware key thus securing and using least bandwidth.
Scanning Module
It facilitates the scanning of paper documents by production level ADF scanners. All scanners based on TWAIN protocol are supported.

Documents can be scanned in various formats like JPEG, BMP, TIFF etc at various resolutions. One sided as well as two sided documents can be scanned by using simplex or duplex options.
Designing Module
With help of this module a user can design a template for a document to be read. This template indicates the position of various fields in a document. Properties of each field-whether it is a character field, numeric field, address field, handwritten or machine printed field etc-are fed into the module.

Co-ordinates of the photo fields are given in this module. These image fields are later indexed and used according to specific needs of the user. Image is cleaned by automatic 'deskewing' ,'despeckling' and 'box removal' features.
Recognition Module
In this module recognition of the data takes place. This process is completely automated and does not require any human intervention. This module can work for 24 hours non-stop.This module reads data from the images and converts that into editable text format or any other format that is required by the user. Recognition module can fix reference points from rectangle, blob, timeline or machine printed characters. Recognition is done by using highly advanced Feature Analysis Algorithms, which provide a high level of accuracy.

The recognition engine combined with our proprietary S2M engine ensures accuracy levels as high as 98%* (* Benchmarking done on PTUCET 2003). Feature richness enables this engine to recognize even unstructured forms.
Proof Reading Module
This is the only module where output is limited by human productivity. Special attention has been paid to make this module user friendly and 'time saving'. Any mistake that might have crept in during recognition can be rectified in this module. S2M engine makes it possible to replace one character with other by a single mouse click.

Keyboard and a virtual keyboard can be put into use for correcting the mistakes. Suspect and unrecognized characters are marked in red.

This module makes it possible for user to send pure data into the database.
Admin Module
In this module users can be created and various activities like scanning, recognition, designing and proofreading can be assigned to different users. Administrator can assign various rights and roles to different users. The Admin module shows the real time progress of each and every batch and users currently logged into Autorec. This enables the administrator to calculate the time needed for the completion of a job.