Stimulation Initiative for European Neural Applications
Esprit Project 9811
Neural OCR Personnel Information Processing at Madrid's Delegacion
Provincial de Educacion
Company Background
Spain's Ministry of Education has more than 300.000 teachers and
educators, distributed throughout the 50 provinces of Spain. On each
one of them, the Delegacion Provincial de Educacion oversees that
province's primary and secondary education, and every year teachers
assignment is obviously a very important task. Madrid's Delegation is
the largest of Spain.
Description of the problem
Given its very high number of teachers, Madrid's Delegacion Provincial
de Educacion has a rather large number of annual teachers reassignments to
new schools, usually involving about 7.000 persons. For obvious
reasons, those reassignments have to be made once the school year has
finished at the end of June. In order to ensure the prompt
knowledge of a teacher's new assignment, the entire process has
to be finished shortly thereafter. This gives the Delegacion
personnel about two weeks to have it concluded, and since every teacher
can request a large number of school destinations (about 150) in
preference order, this implies a very high information processing peak
load. A system that allows the easy incorporation of those assignment
requirements to the Delegacion databases is thus very important for
the successful resolution of a process with significant personal
implications for the persons involved.
Neural Network Techniques application
Neural OCR is becoming the mainstream way of optically recognising
printed material. In the case of Madrid's Delegacion Provincial de
Educacion the PC-based, off-the-shelf system RAMA (developed by the
Instituto de Ingenieria del Conocimiento-IIC) is particularly well
suited to its demands. Besides its excellent OCR capabilities, RAMA
makes possible the definition and redesign of recognition masks
especially tailored for specific forms. In fact, RAMA is used at the
Delegacion for some other tasks such as annual inventory review.
This, together with the name and surname dictionaries, its capacities
to incorporate validation checks and the possibility of incorporating
user defined dictionaries, gives to RAMA a very high versatility for
an organisation such as Madrid's Delegacion that has to deal with
different forms for different tasks.
Benefits
RAMA ensures a prompt and precise processing of the basic information
that Madrid's Delegacion Provincial de Educacion has deemed
interesting for its teacher reassignment process. Besides the modest
hardware requirements, RAMA's ability to handle separately its
scanning, recognition and verification processes allows the
Delegacion to adequately meet its high peak loads.
Generalisation
The OCR application described here is typical of the OCR requirements
of massive personnel information processing. They need a moderate
cost system with modest hardware requirements (a PC and a lower
segment scanner), and the ability to separately process document
scanning, recognition and verification. As the application at Madrid's
Delegacion Provincial de Educacion shows, RAMA is an ideal choice for
these companies.
Contact person
Gonzalo Alvarez-Bueno, Ministerio de Educacion y Ciencias, c/ Vitruvio,=
Madrid