|
|
|---|
Fusion provides organisations with the core functionality they need and at a fraction of the price of more complex systems. Fusion is reshaping the way in which electronic and paper documents are being created, located and managed. Tackling the very first stage in the lifecycle of information, its capture, Fusion provides an on-ramp to enterprise content management (ECM), document management (DM) and workflow systems; feeding them with high quality, validated data and document images.
Formic's development philosophy is to give you the greatest possible payback on your investment in the shortest possible time. We do this by driving down the costs that arise from any subsequent customisation. Fusion software can be put to work immediately and does not require lengthy engagements by specialist programmers to get it working the way you want and keep it working that way as your needs change. What’s more, because all three Fusion systems are built on a common platform, it’s easy to upgrade at any time and without the high costs associated with swapping out systems.
Fusion software is designed to provide a complete information capture solution, seamlessly integrating inbound information from multiple communication channels. Information enters an organisation in various ways: -
| DOCUMENT FORMAT | COMMUNICATION CHANNEL |
|---|---|
| Paper forms and documents | Remote faxes; production scanners; local scanners |
| Web-based forms | Internet; Intranet; Extranet |
| Wireless mobile device forms | Tablet PCs |
Each communication channel has its strengths and weaknesses and no one method of capturing data can be considered ideal in all situations for all respondents. Paper has a great many strengths and will be with us for the foreseeable future. However, with its advantages of speedy transfer, there is an increasing demand for structured information capture on the web and mobile devices. A successful organisation will open up all communication channels for gathering, integrating and validating this data in the same manner to ensure complete data integrity.
Documents can vary enormously in their ability to be read automatically. Processing structured forms offers faster processing times and requires less manual intervention and therefore lend themselves best to automation. Processing semi-structured documents such as invoices, purchase orders and account statements requires more intelligent document recognition technology. As a general rule, semi-structured documents are more complex information capture projects and result in longer processing times and increased costs.
So how exactly does Fusion read people's handwriting? There are different types of handwriting recognition, each requiring a specific technology that delivers varying levels of automation.
There are 5 key categories to consider, the technology behind each of which is progressively more advanced.
Tick Box Recognition
Where marks in tick boxes are read and validated against rule-sets. Crossed out boxes and pen marks in boxes can be ignored. Produces automatic reading rates of up to 100%.
Segmented Image Marking
Marks are made on a picture and picked up by the software according to a preset grid with each location corresponding to an entry database. For instance, a mark on the front right arm on the image shown could input an entry of FrRA in a database. Produces automatic reading rates of up to 95%.
Constrained Text
Upper or lower case handwritten characters appear separated by boxes or combs. Words can be compared against look-up tables to improve accuracy. A reliable and proven technology that delivers automatic reading rates as high as 95%.
Unconstrained Text
Where upper or lower case handwritten characters are not separated. This is more difficult for Intelligent Character Recognition (ICR) engines to 'read' and words are more likely to be compared against look-up tables to improve accuracy. Dependant on the application automatic rates are between 30% and 85% efficient.
Cursive Handwriting
Recognition of joined-up handwriting is more suitable for short pieces of text of limited variation. An example would be reading the value wording of a cheque, as there are a limited number of likely words that would be written. Due to the technical complexity, full page cursive handwriting recognition is not yet available. With the right application, cursive handwriting can be up to 80% automatic.
