Automatic recognition of handwritten numerical strings: a recognition and verification strategy

Oliveira, Luiz S.; Sabourin, Robert; Bortolozzi, Flávio; Suen, Ching Y.

doi:10.1109/tpami.2002.1046154

Cited by 185 publications

(95 citation statements)

References 28 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…In this initial stage of ongoing research, by our oblivious method of compression-based clustering to supply a kernel for an SVM classifier, we achieved a handwritten single decimal digit recognition accuracy of 85%. The current state-of-the-art for this problem, after half a century of interactive feature-driven classification research, in the upper ninety % level [32,14]. All experiments are bench marked on the standard NIST Special Data Base 19 (optical character recognition database).…”

Section: Optical Character Recognitionmentioning

confidence: 99%

Clustering by Compression

Cilibrasi¹,

Vitányi

2005

IEEE Trans. Inform. Theory

922

834

View full text Add to dashboard Cite

We present a new method for clustering based on compression. The method doesn't use subject-specific features or background knowledge, and works as follows: First, we determine a universal similarity distance, the normalized compression distance or NCD, computed from the lengths of compressed data files (singly and in pairwise concatenation). Second, we apply a hierarchical clustering method. The NCD is universal in that it is not restricted to a specific application area, and works across application area boundaries. A theoretical precursor, the normalized information distance, co-developed by one of the authors, is provably optimal in the sense that it minorizes every computable normalized metric that satisfies a certain density requirement. However, the optimality comes at the price of using the non-computable notion of Kolmogorov complexity. We propose precise notions of similarity metric, normal compressor, and show that the NCD based on a normal compressor is a similarity metric that approximates optimality. To extract a hierarchy of clusters from the distance matrix, we determine a dendrogram (binary tree) by a new quartet method and a fast heuristic to implement it. The method is implemented and available as public software, and is robust under choice of different compressors. To substantiate our claims of universality and robustness, we report evidence of successful application in areas as diverse as genomics, virology, languages, literature, music, handwritten digits, astronomy, and combinations of objects from completely different domains, using statistical, dictionary, and block sorting compressors. In genomics we presented new evidence for major questions in Mammalian evolution, based on whole-mitochondrial genomic analysis: the Eutherian orders and the Marsupionta hypothesis against the Theria hypothesis.

show abstract

Section: Optical Character Recognitionmentioning

confidence: 99%

Clustering by Compression

Cilibrasi¹,

Vitányi

2005

IEEE Trans. Inform. Theory

922

834

View full text Add to dashboard Cite

show abstract

“…Finally, the generalization performance of C * j is measured using a test dataset (G). We employ the representation proposed by Oliveira et al [8], which is composed of 132 features. Table 3 lists important information about the database and the partitions used to compose the four separate sets.…”

Section: Methodsmentioning

confidence: 99%

Pareto analysis for the selection of classifier ensembles

Santos

Sabourin

Maupin

2008

Proceedings of the 10th Annual Conference on Genetic and Evolutionary Computation

View full text Add to dashboard Cite

The overproduce-and-choose strategy involves the generation of an initial large pool of candidate classifiers and it is intended to test different candidate ensembles in order to select the best performing solution. The ensemble's error rate, ensemble size and diversity measures are the most frequent search criteria employed to guide this selection. By applying the error rate, we may accomplish the main objective in Pattern Recognition and Machine Learning, which is to find high-performance predictors. In terms of ensemble size, the hope is to increase the recognition rate while minimizing the number of classifiers in order to meet both the performance and low ensemble size requirements. Finally, ensembles can be more accurate than individual classifiers only when classifier members present diversity among themselves. In this paper we apply two Pareto front spread quality measures to analyze the relationship between the three main search criteria used in the overproduce-and-choose strategy. Experimental results conducted demonstrate that the combination of ensemble size and diversity does not produce conflicting multi-objective optimization problems. Moreover, we cannot decrease the generalization error rate by combining this pair of search criteria. However, when the error rate is combined with diversity or the ensemble size, we found that these measures are conflicting objective functions and that the performances of the solutions are much higher.

show abstract

“…Document Text retrieval without using optical character recognition also can be performed [21] along with the minimization in the error rate [22]. Some shapes can be used [23] and verification of recognized result can be performed for enhancing the accuracy [24].…”

Section: Related Workmentioning

confidence: 99%

Marathi Numeral Identification System in Devanagari Script Using 1D Discrete Cosine Transform

Vaidya¹,

Joshi²,

Bhalerao³

2017

IJIES

View full text Add to dashboard Cite

Recognition of handwritten Marathi character/digits is comparatively a tough task as compared to English. It has several types of applications including the postal code reading and sorting, banks check amount processing. In this paper a novel feature extraction and selection method is proposed for the recognition of isolated handwritten Marathi numbers which is based on one dimensional Discrete Cosine Transform (DCT) algorithm. The scanned document is pre-processed and segmented to create isolated numerals. Features for each numeral can be calculated after normalizing the numeral image to 32 × 32 size. Based on these reduced features, the numerals are classified into appropriate groups. Neural network is used for classification of numerals based on the extracted features. The results of proposed method are compared with the results obtained using Discrete Wavelet Transform and zonal discrete cosine transform (ZDCT). The proposed approach gives improved results as compared to zonal DCT and DWT method. Experimental results shows accuracy observed for the method is 90.30% with normalized numeral image of size 32 × 32.

show abstract

Automatic recognition of handwritten numerical strings: a recognition and verification strategy

Cited by 185 publications

References 28 publications

Clustering by Compression

Clustering by Compression

Pareto analysis for the selection of classifier ensembles

Marathi Numeral Identification System in Devanagari Script Using 1D Discrete Cosine Transform

Contact Info

Product

Resources

About