Offline recognition of unconstrained handwritten texts using HMMs and statistical language models

Abstract

This paper presents a system for the offline recognition of large vocabulary unconstrained handwritten texts. The only assumption made about the data is that it is written in English. This allows the application of Statistical Language Models in order to improve the performance of our system. Several experiments have been performed using both single and multiple writer data. Lexica of variable size (from 10,000 to 50,000 words) have been used. The use of language models is shown to improve the accuracy of the system (when the lexicon contains 50,000 words, error rate is reduced by ∼50% for single writer data and by ∼25% for multiple writer data). Our approach is described in detail and compared with other methods presented in the literature to deal with the same problem. An experimental setup to correctly deal with unconstrained text recognition is proposed.

Interactive handwriting recognition with limited user effort

“…This framework also constitutes a successful approach for CAT in HTR [21,23,29]. Traditionally, as stated in [5], the task of HTR can be introduced from a statistical point of view as follows. Given a sequence of feature vectors x = x 1 , · · · , x T = x T 1 representing a text line image, the recognition task can be understood as the search for the sequence of words w that maximises the posterior probability:…”

Towards Unsupervised Learning for Handwriting Recognition

“…Our approach fits well into the standard Hidden Markov Model (HMM) framework based on the sliding window [17][18] [19]. The same concept can be extended to problems such as speech recognition or sign language recognition.…”

Recognition of off-line printed Arabic text using Hidden Markov Models

“…for Arabic text recognition [18][19][20][21][22][23][24][25][26][27][28][29][30][31][32]. HMM offer several advantages.…”

“…Some researchers used HMM for handwriting word recognition [18,19,27,28,31]. Other researchers used it for text recognition [22,23,30,32]. HMM was used for off-line Arabic handwritten digit recognition [33,34] and for character recognition in [30,35].…”

“…In this work, we followed the sliding window principle used by many researcher to extract the features to be used with the HMM [18,20,21,23,25,47]. In our work we employed the sliding window technique used with HMM in speech recognition.…”