Kaltenmeien A, Caesar T, Gloger JM, Mandler E. Sophisticated topology of hidden markov models for cursive script recognition. Proc International Conference on Document Analysis and Recognition, Tsukuba, Japan, 1993; 139--142  Home/Search   Document Not in Database   Summary   Related Articles   Check

....improve the recognition speed of the LVHWR system without changing the recognition rate significantly. 1 Introduction In spite of recent advances in the field of handwriting recognition, few early studies have addressed the problem of large vocabulary off line handwritten word recognition [1] [2] 3] The most frequent simplification has been a pre selection of possible candidate words before the recognition based on other sources of knowledge [4] The majority of works have focused on improving the accuracy of small vocabulary systems while the speed is not taken into account. In ....

Kaltenmeier A, Caesar T, Gloger J M, Mandler E. Sophisticated topology of hidden Markov models for cursive script recognition. 2nd ICDAR, Tsukuba Science City, Japan, 1993, pp 139--142.

....last years handwriting recognition has become an intensive research topic. While the first systems read segmented characters [1] later efforts aimed at the recognition of cursively handwritten words [2] Only a short time ago the first systems appeared which are able to read sequences of words [3, 4, 5]. But these systems operate in small, very specific domains. Only very few systems are known which address the domain of general handwritten text recognition. Among these systems, two fundamentally different approaches have can be observed. In [6, 7] lines of text are segmented into individual ....

A. Kaltenmeier, T. Caesar, J.M. Gloger, and E. Mandler. Sophisticated topology of hidden Markov models for cursive script recognition. In Proc. of the 2nd Int. Conf. on Document Analysis and Recognition, Tsukuba Science City, Japan, pages 139--142, 1993.

....of intensive research. While the first systems read segmented characters, later systems aimed at the recognition of cursively handwritten words. Only short time ago the first systems appeared which are able to read sequences of words. Typical applications of word sequence recognition are address [5] or check reading [6, 2] where recognizers operate in a small, specific domain. At the moment only very few systems are known which address the domain of free text recognition [7] Typically, these systems segment the text into words. But as it is known from the field of continous speech ....

A. Kaltenmeier, T. Caesar, J. Gloger, and E. Mandler. Sophisticated topology of hidden markov models for cursive script recognition. In Proc. of the Second Int. Conf. on Document Analysis and Recognition, pages 139--142, 1993.

....the same accuracy and consuming a reasonable amount of memory. 1 Introduction Off line recognition of handwritten words is a challenging task due to the high variability and uncertainty of human writing. Several proposals to solve this problem have been presented recently [1] 2] 3] 4] [5] [11] The majorityofthe state of art systems have some constraints during the recognition task. One of the most common constraints is the limitation of the size of the lexicon that a system can deal with. Open vocabulary systems, that is, systems that do not rely on a lexicon during the ....

....how to manage the search complexities in a large vocabulary, especially in real time applications, poses a serious challenge to the researchers. Several techniques such as breadth first, depth first and best first search have been employed to search and recover words from lexical trees [5] [6] 7] Nevertheless, our problem is slightly more complex than just searching and recovering a word from the lexical tree. In fact, we need to traverse the whole tree since we do not know in advance which word we are looking for because the answer is not a single node but a set of linked nodes ....

A. Kaltenmeier, T. Caesar, J. M. Gloger, and E. Mandler. Sophisticated Topology of Hidden Markov Models for Cursive Script Recognition. In 2nd International Conference on Document Analysis and Recognition, Tsukuba Science City, Japan, October 1993.

.... can be categorized into two groups, those for recognizing connected character strings with fixed number of characters [1, 2, 3, 4] and those for recognizing character strings without knowing the length of each string ( unfixed ) where the segmentation and recognition are performed interactively [5, 6, 7, 8, 9, 10, 11, 12]. Cheriet et al. presented a region based background analysis algorithm to find a married pair of background valleys in order to separate a connected digit string [1] On the other hand, a context directed hierarchical algorithm was proposed by Shridhar and Badreldin to separate connected 2 digit ....

....cut links where the weights were trained using a genetic algorithm. And a integrate segmentation algorithm was proposed by Pervez et al. [13] Most of them have a good performance. In the second group, Hidden Markov Model (HMM) a segmentation free approach, is applied in most applications [5, 6, 7, 8, 11]. An HMM is a doubly stochastic process with an underlying stochastic process which is not observable but can be observed through another set of stochastic processes that produce the sequence of symbols [14] It does not model the whole pattern or shape as a single feature vector. Instead, the HMM ....

A. Kaltenmeier, T. Caesar, J. M. Gloger, and E. Mandler. Sophisticated topology of hidden Markov models for cursive script. In Proceedings of ICASSP'93, pages 139 -- 142, 1993.

....of our system in recognizing documents of poor quality by presenting results on faxed English data. A number of research efforts made use of HMMs for off line printed and handwriting recognition, but recognition was always performed on a single language. Our approach is most similar to those of [1,7,9,11,12,17] in that we are extracting features from thin slices of the line image, which is the key to making the recognition system language independent. However, often additional feature extraction steps are taken, as in Elms and Illingworth [9] who also extract features from horizontal slices, which ....

....system (developed for recognition on printed Roman characters) inappropriate for languages with connected script. Aas and Eikvil [1] used a bounding box around each word to be recognized and extracted features from vertical thin slices to perform recognition on a single printed Roman font. Kornai [11] also used features extracted from vertical thin slices to perform recognition on handwritten addresses from the CEDAR corpus. There has been little work in using HMMs for the recognition of Arabic script [2] Allam [3] used contour tracing to locate groups of connected characters; the recognition ....

A. Kaltenmeier, T. Caesar, J.M. Gloger, and E. Mandler, "Sophisticated topology of hidden Markov models for cursive script recognition," Proc. Int. Conf. Document Analysis and Recognition, Tsukuba City, Japan, 139-142, 1993.

....adding new words to the lexicon. In another approach, an HMM character subcharacter model method is used. It reduces the amount of parameter set and increases efficiency on modeling. For word recognition, word models are constructed by linking corresponding character sub character models [7] [8]. However, it needs large search space for large vocabulary lexicon. As an another usage, HMM character sub character models are used to make lattice. Lattice search, level building method [9] or Viterbi method [10] is applied for recognition. In this method, lexicon size can be increased ....

A. Kaltenmeier, T. Caesar, J. M. Gloger, E. Mandler, "Sophisticated Topology of Hidden Markov Models for Cursive Script Recognition," Proc. of 2nd Int. Conf. on Document Analysis and Recognition, Tsukuba Science City, Japan, pp.139--142, Oct. 1993

No context found.

Kaltenmeien A, Caesar T, Gloger JM, Mandler E. Sophisticated topology of hidden markov models for cursive script recognition. Proc International Conference on Document Analysis and Recognition, Tsukuba, Japan, 1993; 139--142

No context found.

A. Kaltenmeier, T. Caesar, J. M. Gloger, and E. Mandler. Sophisticated topology of hidden markov models for cursive script recognition. In Proc. International Conference on Document Analysis and Recognition, pages 139-- 142, Tsukuba, Japan, 1993.

No context found.

Kaltenmeien A, Caesar T, Gloger JM, Mandler E. Sophisticated topology of hidden markov models for cursive script recognition. Proc International Conference on Document Analysis and Recognition, Tsukuba, Japan, 1993; 139--142

No context found.

A. Kaltenmeier, T. Caesar, J. M. Gloger, and E. Mandler. Sophisticated topology of hidden markov models for cursive script recognition. In Proc. International Conference on Document Analysis and Recognition, pages 139-- 142, Tsukuba, Japan, 1993.

No context found.

A. Kaltenmeier, T. Caesar, J. M. Gloger, and E. Mandler. Sophisticated Topology of Hidden Markov Models for Cursive Script Recognition. In 2nd International Conference on Document Analysis and Recognition, Tsukuba Science City, Japan, October 1993.

No context found.

A. Kaltenmeier, T. Caesar, J. M. Gloger, and E. Mandler, Sophisticated Topology of Hidden Markov Models for Cursive Script Recognition, 2nd International Conference on Document Analysis and Recognition, Tsukuba Science City, Japan, Oct 20-22, 1993.

Online articles have much greater impact   More about CiteSeer.IST   Add search form to your site   Submit documents   Feedback  

CiteSeer.IST - Copyright Penn State and NEC