TY - BOOK
T1 - Handwritten line detection via an em algorithm
AU - Cruz, Francisco
AU - Terrades, Oriol Ramos
PY - 2013
Y1 - 2013
N2 - In this paper we present a handwritten line segmentation method devised to work on documents composed of several paragraphs with multiple line orientations. The method is based on a variation of the EM algorithm for the estimation of a set of regression lines between the connected components that compose the image. We evaluated our method on the ICDAR2009 handwriting segmentation contest dataset with promising results that overcome most of the presented methods. In addition, we prove the usability of the presented method by performing line segmentation on the George Washington database obtaining encouraging results.
AB - In this paper we present a handwritten line segmentation method devised to work on documents composed of several paragraphs with multiple line orientations. The method is based on a variation of the EM algorithm for the estimation of a set of regression lines between the connected components that compose the image. We evaluated our method on the ICDAR2009 handwriting segmentation contest dataset with promising results that overcome most of the presented methods. In addition, we prove the usability of the presented method by performing line segmentation on the George Washington database obtaining encouraging results.
KW - Expectation-Maximization algorithm
KW - Handwritten Line Segmentation
KW - Linear regression
UR - http://www.scopus.com/inward/record.url?scp=84889566404&partnerID=8YFLogxK
U2 - 10.1109/ICDAR.2013.147
DO - 10.1109/ICDAR.2013.147
M3 - Proceeding
AN - SCOPUS:84889566404
T3 - Proceedings of the International Conference on Document Analysis and Recognition, ICDAR
BT - Handwritten line detection via an em algorithm
ER -