Comparison of Bernoulli and Gaussian HMMs using a vertical repositioning technique for off-line handwriting recognition

—In this paper a vertical repositioning method based on the center of gravity is investigated for handwriting recognition systems and evaluated on databases containing Arabic and French handwriting. Experiments show that vertical distortion in images has a large impact on the performance of HMM base...

Full description

Bibliographic Details
Authors: Doetsch, Patrick, Hamdani, Mahdi, Ney, Hermann, Andrés Ferrer, Jesús, Giménez Pastor, Adrián, Juan, Alfons|||0000-0002-9984-4072
Format: book part
Publication Date:2012
Country:España
Institution:Universitat Politècnica de València (UPV)
Repository:RiuNet. Repositorio Institucional de la Universitat Politécnica de Valéncia
Language:English
OAI Identifier:oai:riunet.upv.es:10251/50129
Online Access:https://riunet.upv.es/handle/10251/50129
Access Level:Open access
Keyword:Handwriting recognition
Vertical distortion
Center of gravity
Recurrent neural networks
Bernoulli HMMs
ESTADISTICA E INVESTIGACION OPERATIVA
LENGUAJES Y SISTEMAS INFORMATICOS
Description
Summary:—In this paper a vertical repositioning method based on the center of gravity is investigated for handwriting recognition systems and evaluated on databases containing Arabic and French handwriting. Experiments show that vertical distortion in images has a large impact on the performance of HMM based handwriting recognition systems. Recently good results were obtained with Bernoulli HMMs (BHMMs) using a preprocessing with vertical repositioning of binarized images. In order to isolate the effect of the preprocessing from the BHMM model, experiments were conducted with Gaussian HMMs and the LSTM-RNN tandem HMM approach with relative improvements of 33% WER on the Arabic and up to 62% on the French database.