Strategies for handwritten words recognition using hidden Markov models

Gilloux, M.; Leroux, Manuel; Bertille, J.-M.

doi:10.1109/icdar.1993.395727

Cited by 18 publications

(13 citation statements)

References 4 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…Table 2 summarizes all segmentation-based methods according to the different categories they were 2 There was only one piece of evidence we ran into during our survey of word recognition systems where a segmentation algorithm was used in a small static lexicon environment. Moreover, we found this single method by Gilloux et al, [32][33][34], to have a lot in common with segmentation-free methods despite the segmentation algorithm that it was integrated with. Further discussion of this exception was given in the previous section.…”

Section: Segmentation-based Methodsmentioning

confidence: 99%

“…All methods by Guillevic et al, [39], Saon et al, [70], and Gilloux et al, [32][33][34], represent a word model as a chain of n identical sub-HMMs (see an example in Fig. 5), where n is the most probable length of an observation sequence obtained from the training samples.…”

Section: Hmmsmentioning

confidence: 99%

“…A null transition from the initial state to that of the next sub-HMM [32]. This one was replaced by a normal transition that bypasses the following sub-HMM in the chain in [33,34]. 2.…”

Section: Hmmsmentioning

confidence: 99%

“…The final recognition score of a word model is therefore p(o|w) × p(w), found by the Viterbi algorithm [85,26]. In other systems by Gilloux et al, [30,31,33,34] the same elementary sub-HMMs were used to construct letter models. All the above-mentioned methods by Gilloux et al, used a segmentation algorithm to produce observations that were symbols describing each segment.…”

Section: Hmmsmentioning

confidence: 99%

See 3 more Smart Citations

Offline cursive script word recognition ? a survey

Steinherz

Rivlin

Intrator

1999

International Journal on Document Analysis and Recognition

View full text Add to dashboard Cite

Abstract.We review the field of offline cursive word recognition. We mainly deal with the various methods that were proposed to realize the core of recognition in a word recognition system. These methods are discussed in view of the two most important properties of such a system: the size and nature of the lexicon involved, and whether or not a segmentation stage is present. We classify the field into three categories: segmentation-free methods, which compare a sequence of observations derived from a word image with similar references of words in the lexicon; segmentation-based methods, that look for the best match between consecutive sequences of primitive segments and letters of a possible word; and the perception-oriented approach, that relates to methods that perform a human-like reading technique, in which anchor features found all over the word are used to bootstrap a few candidates for a final evaluation phase.

show abstract

Section: Segmentation-based Methodsmentioning

confidence: 99%

Section: Hmmsmentioning

confidence: 99%

“…A null transition from the initial state to that of the next sub-HMM [32]. This one was replaced by a normal transition that bypasses the following sub-HMM in the chain in [33,34]. 2.…”

Section: Hmmsmentioning

confidence: 99%

Section: Hmmsmentioning

confidence: 99%

See 2 more Smart Citations

Offline cursive script word recognition ? a survey

Steinherz

Rivlin

Intrator

1999

International Journal on Document Analysis and Recognition

View full text Add to dashboard Cite

show abstract

“…The random variables whose joint probabilities must be estimated are letter, word, or part-of-speech labels. Linguistic context is always order-dependent, and therefore often modeled with transition frequencies in Markov Chains, Hidden Markov Models, and Markov Random Fields [16,17,18,19,20,21,22,23,24,25,26,27,28]. Linguistic variables are usually assumed to be independent of character shape, even though titles and headings in large or bold type have a different language structure than plain text.…”

Section: Language Modelsmentioning

confidence: 99%

Adaptive and Interactive Approaches to Document Analysis

Nagy¹,

Veeramachaneni

2008

Machine Learning in Document Analysis and Recognition

View full text Add to dashboard Cite

Summary. This chapter explores three aspects of learning in document analysis: (1) field classification, (2) interactive recognition, and (3) portable and networked applications. Context in document classification conventionally refers to language context, i.e., deterministic or statistical constraints on the sequence of letters in syllables or words, and on the sequence of words in phrases or sentences. We show how to exploit other types of statistical dependence, specifically the dependence between the shape features of several patterns due to the common source of the patterns within a field or a document. This type of dependence leads to field classification, where the features of some patterns may reveal useful information about the features of other patterns from the same source but not necessarily from the same class. We explore the relationship between field classification and the older concepts of unsupervised learning and adaptation. Human interaction is often more effective interspersed with algorithmic processes than only before or after the automated parts of the process. We develop a taxonomy for interaction during training and testing, and show how either human-initiated and machine-initiated interaction can lead to human and machine learning. In a section on new technologies, we discuss how new cameras and displays, web-wide access, interoperability, and essentially unlimited storage provide fertile new approaches to document analysis.

show abstract