Con esta herramienta te facilitamos un acceso a todas las ofertas y demandas de tecnología europeas y a búsquedas de socios para participar en propuestas europeas de I+D publicadas en la red Enterprise Europe Network, pudiendo filtrar los resultados para facilitar las búsquedas más acordes con tus necesidades.

¿Quieres recibir estos listados de oportunidades de colaboración en tu correo de forma periódica y personalizada? Date de alta en nuestro Boletín

Para optimizar los resultados de la búsqueda, se recomienda utilizar términos en inglés.

Empresa austríaca busca una herramienta de análisis de archivos gráficos/PDF para conversión en formato de lectura electrónica

Resumen

Tipo:
Demanda Tecnológica
Referencia:
TRAT20190926002
Publicado:
15/10/2019
Caducidad:
15/10/2020
Resumen:
Una pequeña empresa austríaca de ingeniería busca una solución para extraer elementos de un documento complejamente estructurado (principalmente PDF pero también Word, Excel, PowerPoint y .jpg) y convertirlo en un formato de lectura electrónica. Los elementos incluyen texto, gráficos, dibujos, tablas, encabezados y módulos de información. La solución debe ofrecer una precisión del 90-99% dependiendo de la complejidad del texto. También está interesada en versiones futuras de la solución que puedan gestionar tablas de contenidos, caracteres no latinos (chino, japonés y coreano) e imágenes de fondo complejas. La empresa busca socios con el fin de establecer acuerdos de licencia, comercialización con asistencia técnica o investigación.

Details

Tittle:
PDF/graphic file analysis tool sought for conversion to machine-readable format
Summary:
A small Austrian engineering company seeks a solution, which can extract the elements of a given complexly structured document (primarily PDF but also other formats) and convert them into a machine-readable form. Elements include text, graphics, drawings, tables, headlines and info-boxes. Very high accuracy is needed (90-99% correct readings). The company seeks partners for licence agreement, commercial agreement with technical assistance or research cooperation agreement.
Description:
An Austrian engineering firm is looking for a solution to facilitate information processing and the analysis of data particularly in PDF and graphic format.

When for example researching a topic online, most information is contained in free formatted documents, which contain a number of different elements. In order to analyse the documents, they first need to be broken down into formal elements such as text, images, headings, tables etc. and the logical links need to be retained. Only then can the contents be interpreted by specialised, external services, e.g. NLU (natural language understanding) or image recognition.

While there are a number of solutions currently available, the one major shortcoming is their insufficient extraction hit rate. The offered solution must deliver accuracy of 90-99% depending on the text complexity. Current solutions are also very focused on plain text. Context information (headings, text placement, to which graphic does it refer), tables, multiple columns of the text, graphics, images, captions or comments cannot be reliably extracted in context.

The Austrian SME is looking for a solution to handle input primarily in the form of good quality PDF scans. However, a tool that can also deal with Word, Excel, PowerPoint and .jpg formats would be ideal. It is important that the solution can handle complex content with headlines, sublines, different text bodies, columns of text, drawings, graphics, tables, nesting tables, info boxes (text) and sometimes background pictures.

The tool should cover the following processing steps:

· breaking down a given document into the elements mentioned above. Context-information should be retained (headline of text body, page number, elements close to each other etc.);
· converting text, numbers and tables into machine-readable formats;
· displaying confidentiality-index of each separate element described in the section "technical specification or expertise sought"
· outputting the results in file systems or databases.

The technical specifications for the initial solution are detailed in the section "technical specification or expertise sought". Future versions of the solution, which can handle tables of contents, non-Latin characters (Chinese, Japanese, Korean) and more complex background images are also of interest.

The ideal cooperation types include:

- a licence agreement where a vendor is willing to sell a licence for a suitable solution;
- a commercial agreement with technical assitance with a developer who can deliver the required framework with complete knowledge transfer. In this case, initial training should be provided by the developer to an extent that (almost) no further training is required;
- a research cooperation agreement, where the Austrian company would work together with another party to develop the technology, should a suitable company or research team be found.
Technical Specification or Expertise Sought:
Once the input has been processed, the output must include the following:

· complete data as per the original document
· text fields and paragraphs in ASCII (American Standard Code for Information Interchange) , RTF (rich text format) etc. annotated with context (link to a heading or the depicted image, document page number etc.)
· document heading
· sub text of the heading or chapter title
· text of info boxes (possibly denoted with a frame and displayed on a grey background)
· chapter/paragraph titles annotated with respect to the body text
· images, drawings, graphics annotated with respect to the image information
· image information, image heading/caption, info blocks which refer to the image
· content of tables and nesting tables in a machine-readable format
· page numbers
· accurate sequences
· management of cross-page elements and multiple columns
· accuracy of 90-99% (invariably dependent on the quality of the document, background colour and selection of font, font size and special characters)

Partner sought

Type and Role of Partner Sought:
Suitable partners include companies as vendors of such a service (cloud or on site) or a specialised developer active in e.g. text recognition or a university with a proven track record in this field. Partnerships considered include:

- a licence agreement where a vendor is willing to sell a licence for a suitable solution;
- a commercial agreement with technical assitance with a developer who can deliver the required framework with complete knowledge transfer. In this case, initial training should be provided by the developer to an extent that (almost) no further training is required;
- a research cooperation agreement, where the Austrian company would work together with another party to develop the technology, should a suitable company or research team be found.

Client

Type and Size of Client:
Industry SME <= 10
Already Engaged in Trans-National Cooperation:
No
Languages Spoken:
English
German