Periodic Reporting for period 1 - HandBiblio (HandBiblio – Semantic Search Engine for Handwritten Documents - Enabling an effective andautomatic access to contents of handwritten documents, stored as digital images.)
Reporting period: 2017-12-01 to 2018-04-30
In particular handwriting is one of the most ancient means for communicate our ideas and concepts and that’s the reason why the digitalization of paper documents is a mandatory process to preserve and maintain the knowledge contained into handwritten documents.
At N.I.Te we are really convinced that as long as the human being will have his hands, he will use them to write something, so that somebody can read it.
Nowadays, the need for an automatic processing of handwritten texts is of outmost importance for any organization, but whether all computer-based reading systems are able to recognize and acquire typewritten text, they fail when attempting to recognize manual writing in italics, unless very huge computing power and very long training times are employed. For this reason, very often crucial information is lost or altered, procuring a potentially invaluable harm to the organization. Our approach is based on a brand-new idea, which consists not in recognizing individual characters belonging to words, but the whole ""ink strokes,"" regardless of whether they represent individual characters or syllables or even full words and independently of the writer, the alphabet and the language. The result is that required training time is dramatically reduced and the operation is natively independent from the writer, the alphabet and the language: to understand documents handwritten in Mandarin and Arabic language can be as easy as our mother tongue.
Recognition of handwriting characters today is a service carried out through platforms requiring a very time intensive and cost expensive set-up; it is prohibitive for most enterprises and only accessible to those organizations that have to deal with a large amount of data that justify long months working and employing of domain specialists. The recognition of handwriting characters proposed by NITE instead, is an innovative service as for its methods of data acquisition and processing: due to its peculiarity, the system does not require long times for preparation and initial setup.
Our job is building reading machines with Artificial Intelligence algorithms.
Our mission is to become one of the national and international excellence in the ICT sector of Pattern Recognition and Artificial Intelligence. The EC support and funding would be very important in order to reach our financial and strategic objectives.
Our vision is to “Make available and searchable all the knowledge contained in handwritten digital documents of the world.”
1. From a marketing point of view, we have better examined, segmented and measured our market, which is growing strongly worldwide (over 15% a year for the next eight years), and we setup our goal is to conquer a slice of this new market, betting on our innovative technology. A first positioning of our offering (to be considered a starting point for phase 2) has been realised in the segment of ""Offline"" market applications, i.e. applications working on images of existing manuscripts, rather than text acquisition while writing (""Online applications""). Within the Offline segment, the four most promising application fields was identified: Educational & Literature, Enterprise management, Manufacturing, Healthcare. As an objective of phase 2, it is our intention to address each of these four application fields with a specific vertical solution. The business model to apply will be implemented as distribution network based in some European and non European Countries plus an Internet channel, aimed at selling the Vertical Applications on a monthly license fee basis. Tailored solutions projects addressing special customers requirements and an SDK package for developers purposes, will complete the offering portfolio. During phase 1 a first generic communication plan was drafted.
2. From a technical point of view, we have designed and developed a prototype infrastructure, aimed at demonstrating our technology in industrially relevant environment (TRL 6). The best technical HW-SW environments and application scenarios have been selected through a specific assessment activity and three service lines have been envisaged, namely: historical form processing, contemporary form processing and postal form processing. Some Hw equipment such as smart glasses, scanners and Hi-resolution cameras have been tested, in order to check both performances and user friendliness features.
We also had the opportunity to setup technical tests and demos with several public and private organizations and companies, especially with owners of historical handwritten collections that provide us samples of data. The results of these demos were very promising in the sense that we have obtained some evidence of the need of public and private organizations to automatically manage their handwritten digital archives and their needs to invest resources first in the digitalization of their paper collection and then on effective management of these huge digital archives. Some test session was also carried out on non Latin alphabets (mainly Japanese logographic), achieving a first positive results. Work on this topic is currently on going.
3. From an operational point of view, the phase 1 business plan was reviewed and an assessment was carried out concerning the expected impact in terms of turnover, employees, costs and margins. A Risk analysis was also carried out, mainly aimed at facing potential difficulties in organizing and effectively manage large-scale dissemination and exploitation, as well as the difficulties to face a very rapid growth (overtrading). During this step, an analysis concerning the possibility of establishing a Joint Venture or a Spin-Off with other companies operating into the same market segment (external growth) was also explored. In order to achieve this last result, some brokerage activities have been performed, among which the most relevant is the attendance to the event organised by the Business Acceleration Services of European Commission at Hannover Fair. This was a great opportunity for our company to meet several potential partners interested in possible future technological and commercial partnerships."
Thanks to the HandBiblio project, all organizations requiring to manage handwritten documents electronically, will discard manual metadata and transcription currently used in digitization processes, without having to resort to prohibitive solutions in terms of cost, time and processing capacity, regardless of their size and business sector, and whatever their language and alphabet.
We believe the proposed goals are within our reach thanks to an expected strong expansion of the market, but above all thanks to our value proposition: “an innovative technology that ensures to all users the most affordable, efficient, easy to use and comprehensive electronic document management, also reducing time for processing and expertise requirements”.
Our project will last two years and envisages a growth of the company’ staff from the current 7 people to 15 people at the end of the project, to around 50 people after three years from the end of the project.