Anja Brakensiek

Combination of multiple classifiers for handwritten word recognition

Due to large shape variations in human handwriting, recognition accuracy of cursive handwritten word is hardly satisfying using a single classifier. In this paper we introduce a framework to combine results of multiple classifiers and present an intuitive run-time weighted opinion pool combination approach for recognizing cursive handwritten words with a large size vocabulary. The individual classifiers are evaluated run-time dynamically. The final combination is weighted according to their local performance. For an open vocabulary recognition task, we use the ROVER algorithm to combine the different strings of characters provided by each classifier. Experimental results for recognizing cursive handwritten words demonstrate that our new approach achieves better recognition performance and reduces the relative error rate significantly.

Handwritten Address Recognition Using Hidden Markov Models

In this paper several aspects of a recognition system for cursive handwritten German address words (cities and streets) are described. The recognition system is based on Hidden Markov Models (HMMs), whereat the focus is on two main problems: the changes in the writing style depending on time or regional differences and the difficulty to select the correct (complete) dictionary for address reading. The first problem leads to the examination of three different adaptation techniques: Maximum Likelihood (ML), Maximum Likelihood Linear Regression (MLLR) and Scaled Likelihood Linear Regression (SLLR). To handle the second problem language models based on backoff character n-grams are examined to evaluate the performance of an open vocabulary recognition (without dictionary). For both problems the determination of confidence measures (based on the frame-normalized likelihood, a garbage model, a two-best recognition or an unconstrained character decoding) is important, either for an unsupervised adaptation or the detection of out of vocabulary words (OOV). The databases, which are used for recognition, are provided by Siemens Dematic (SD) within the Adaptive READ project.

Confidence measures for an address reading system

In this paper the performance of different confidence measures used for an address recognition system are evaluated. The recognition system for cursive handwritten German address words is based on hidden Markov models (HMMs). It is essential that the structure of the address (name, street, city, country) is known, so that a specific small but complete dictionary can be selected. Upon choosing a wrong dictionary (OOV: out-of-vocabulary) or misrecognizing a word, the recognition result should be rejected by means of the confidence measure. This paper points out two aspects: the comparison of four confidence measures for single words - based on the likelihood, a garbage-model, a two-best recognition or a character decoding - and the comparison of using complete or wrong dictionaries. It is shown that the best confidence measure - the two-best distance - has a quite different behavior using OOV.

Comparing adaptation techniques for on-line handwriting recognition

This paper describes an online handwriting recognition system with focus on adaptation techniques. Our hidden Markov model (HMM)-based recognition system for cursive German script can be adapted to the writing style of a new writer using either a retraining depending on the EM (expectation maximization)-approach or an adaptation according to the MAP (maximum a posteriori) or MLLR (maximum likelihood linear regression)-criterion. The performance of the resulting writer-dependent system increases significantly even if the amount of adaptation data is very small (about 6 words). So this approach is also applicable for online systems in hand-held computers such as PDAs. Special attention was paid to the performance comparison of the different adaptation techniques with the availability of different amounts of adaptation data ranging from a few words tip to 100 words per writer.

Performance evaluation of a new hybrid modeling technique for handwriting recognition using identical on-line and off-line data

The paper deals with the performance evaluation of a novel hybrid approach to large vocabulary cursive handwriting recognition and contains various innovations. 1) It presents the investigation of a new hybrid approach to handwriting recognition, consisting of hidden Markov models (HMMs) and neural networks trained with a special information theory based training criterion. This approach has only been recently introduced successfully to online handwriting recognition and is now investigated for the first time for offline recognition. 2) The hybrid approach is extensively compared to traditional HMM modeling techniques and the superior performance of the new hybrid approach is demonstrated. 3) The data for the comparison has been obtained from a database containing online handwritten data which has been converted to offline data. Therefore, a multiple evaluation has been carried out, incorporating the comparison of different modeling techniques and the additional comparison of each technique for online and offline recognition, using a unique database. The results confirm that online recognition leads to better recognition results due to the dynamic information of the data, but also show that it is possible to obtain recognition rates for offline recognition that are close to the results obtained for online recognition. Furthermore, it can be shown that for both online and offline recognition, the new hybrid approach clearly outperforms the competing traditional HMM techniques. It is also shown that the new hybrid approach yields superior results for the offline recognition of machine printed multifont characters.

Multi-branch and two-pass HMM modeling approaches for off-line cursive handwriting recognition

Because of large shape variations in human handwriting, cursive handwriting recognition remains a challenging task. Usually, the recognition performance depends crucially upon the pre-processing steps, e.g. the word baseline detection and segmentation process. Hidden Markov models (HMMs) have the ability to model similarities and variations among samples of a class. In this paper, we present a multi-branch HMM modeling method and an HMM-based two-pass modeling approach. Whereas the multi-branch HMM method makes the resulting system more robust with word baseline detection, the two-pass recognition approach exploits the segmentation ability of the Viterbi algorithm and creates another HMM set and carries out a second recognition pass. The total performance is enhanced by the combination of the two recognition passes. Experiments recognizing cursive handwritten words with a 30,000-word lexicon have been carried out. The results demonstrate that our novel approaches achieve better recognition performance and reduce the relative error rate significantly.

Handwritten address recognition with open vocabulary using character n-grams

In this paper a recognition system, based on tied-mixture hidden Markov models, for handwritten address words is described, which makes use of a language model that consists of backoff character n-grams. For a dictionary-based recognition system it is essential that the structure of the address (name, street, city) is known. If the single parts of the address cannot be categorized, the used vocabulary is unknown and thus unlimited. The performance of this open vocabulary recognition using n-grams is compared to the use of dictionaries of different sizes. Especially, the confidence of recognition results and the possibility of a useful post-processing are significant advantages of language models.

Unlimited Vocabulary Script Recognition Using Character N-Grams

In this paper a robust Script recognition system is described, which makes use of a language model, that consists of backoff character n-grams. The system is based on Hidden Markov Models (HMMs) using discrete and hybrid modeling techniques, where the latter depends on a vector quantizer trained according to the MMI-criterion (information theory-based neural network). The presented recognition results refer to the SEDAL-database of degraded English documents such as photocopy or fax using no dictionary and a writer-dependent handwritten database of cursive German Script samples. Our resulting system for character recognition yields significantly better recognition results for an unlimited vocabulary using language models.

Comparing normalization and adaptation techniques for online handwriting recognition

In this paper a writer-independent online handwriting recognition system is described comparing the influence of handwriting normalization and adaptation techniques on the recognition performance. Our hidden Markov model-based recognition system for unconstrained German script can be adapted to the writing style of a new writer using different adaptation techniques whereas the impact of preprocessing to normalize the pen-trajectory is examined. The performance of the resulting writer-dependent system increases significantly, even if only a few words are available for adaptation. So this approach is also applicable for online systems in hand-held computers such as PDAs. In addition, the developed normalization techniques are helpful to improve completely writer independent systems. This paper presents the performance comparison of three different adaptation techniques either in a supervised or unsupervised mode, in combination with appropriate normalization methods, with the availability of different amount of adaptation data ranging from only 6 words up to 100 words per writer.

A comparison of character n-grams and dictionaries used for script recognition

In this paper an off-line script recognition system is described, which makes use of a language model, that consists of backoff character n-grams. The performance of this open vocabulary recognition is compared with the use of closed dictionaries. The system is based on Hidden Markov Models (HMMs) using a hybrid modeling technique, which depends on a neural vector quantizer The presented recognition results refer to the SEDAL-database of degraded English documents such as photocopy, or fax and a writer-dependent handwritten database of cursive German script samples. Our resulting system for character recognition yields significantly, better recognition results for an unlimited vocabulary using language models.