Publications

Advanced search

Abstract

Lorenzo Quirós, Carlos D. Martínez-Hinarejos, Alejandro H. Toselli, Enrique Vidal. Interactive Layout Detection. IbPRIA 2017, 2017. pp. 161-168. Springer.

The amounts of ancient documents transcribed by means of Handwritten Text Recognition (HTR) technology have been rising dramatically over the last years. Consequently, the development and enhancement of HTR methods and algorithms have become an important issue in the field, with significant contributions in performance for documents with segmented layout. However, Layout Analysis remains a bottleneck in the development and generalization of HTR technology. In this work a new Interactive-Probabilistic method to obtain document layout is presented. This new method incorporates the user feedback in the Layout Analysis process, in order to provide not just a very accurate layout, but an interactive framework in which user feedback is used to help the system to fix any error.