Overview
Handwritten word spotting is the task of detecting query words in handwritten
document image collections without involving a traditional OCR procedure.
Recently, handwritten word spotting has attracted the attention of the
research community in the field of document image analysis and
recognition since it has been proved being a feasible solution in handwritten documents indexing and retrieval.
We invite all
researchers in the field of handwritten document analysis and recognition to register
for participating in H-WSCO 2013. The general objective of this contest is
to evaluate recent approaches for handwritten
word spotting. A query word (image or text) as well as a set of handwritten document images
are provided as input. The output consists of a subset of the
initial set of images as well as the corresponding bounding box
coordinates for each spotted word instance.
In the context of the competition, a benchmarking dataset will be built along with an evaluation methodology
in order to test and compare recent algorithms for handwritten word
spotting in realistic circumstances. For the preparation of the
participants, a representative portion of the dataset along with its
corresponding ground truth
will be available for experimentation. Datasets will contain
binary handwritten
document images produced by several writers in English language.
An agreement
will be signed by the participants and the organizers in order to
protect the rights of the submitted software. The results of the
competition will be announced during a dedicated ICDAR 2013
session. A
report on the competition will be published in the ICDAR 2013
conference
proceedings.
Note: You may participate in this contest even if you do not plan to attend the ICDAR 2013 conference.
B.Gatos1, G. Louloudis1, N. Stamatopoulos1,
A. Kesidis2, I. Pratikakis3 and K. Zagoris3
1National Center for Scientific Research "Demokritos", Greece
2Technological Educational Institute of Athens, Greece
3Democritus University of Thrace, Greece