An efficient three-stage classifier for handwritten digit recognition

Gorgevik, D.; Cakmakov, Dusan

doi:10.1109/icpr.2004.1333822

Cited by 22 publications

(8 citation statements)

References 6 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The most popular feature extraction methods comprising of projection histogram, or contour profile are approached by many researches to confirm the effective performance as in [3][6] [8]. Firstly, contour profile scheme is introduced in the Fig.…”

Section: B Feature Extractionmentioning

confidence: 99%

Isolated vietnamese handwriting recognition embedded system applied combined feature extraction method

Van

Huu

Hoang

2015

2015 International Conference on Advanced Technologies for Communications (ATC)

View full text Add to dashboard Cite

In this research, an embedded system approaching isolated Vietnamese handwriting upper case recognition is illustrated. Obtained results from experiments have experienced that proposed feature extraction method, a major step of complete character recognition scheme, combining contour profile and projection histogram techniques is superior to others. Furthermore, the full conceptual system performance is evaluated on the embedded system, namely BeagleBoard-xM board, to confirm that recognition probability is up to more 90% over 27000 characters over thirty different persons.

show abstract

Section: B Feature Extractionmentioning

confidence: 99%

Isolated vietnamese handwriting recognition embedded system applied combined feature extraction method

Van

Huu

Hoang

2015

2015 International Conference on Advanced Technologies for Communications (ATC)

View full text Add to dashboard Cite

show abstract

“…One could build a cascade to increase the accuracy [RAHMAN & FAIRHURST 1999], or to increase the speed of the classification system [KAYNAK & ALPAYDIN 1997, PUDIL et al 1992, GIUSTI ET AL. 2002, GORGEVIK & CAKMAKOV 2004, ,FERRI et al 2004, Chellapilla et al 2006a, Chellapilla et al 2006b]. …”

Section: Related Workmentioning

confidence: 99%

“…A cascade could be manually built [GORGEVIK & CAKMAKOV 2004, ], or automatically built [Chellapilla et al 2006a, Chellapilla et al 2006b]. The degree of cascade building automation differs.…”

Section: -Manual Versus Automatic Building Of Cascadesmentioning

confidence: 99%

A Model-Based Approach for Building Optimum Classification Cascades

El-Sherif

Abdelazeem

2010

Pattern Recognition Recent Advances

View full text Add to dashboard Cite

For a long time, the main concern of the pattern recognition research community has been to achieve high accuracy. For example, in the area of Multi-Classifier Systems (MCS), researchers have developed very powerful techniques that combine large number of classifiers through complex combination schemes to achieve satisfactory accuracies. However, this was at the expense of complexity. Even powerful single classifiers (e.g. SVM) have very high complexity. This might give the impression that high accuracies could not be achieved without sacrificing recognition time. Classification cascades are relatively neglected because speed is usually considered a secondary issue by researchers in pattern recognition field. This fact is going to change in the near future because as the world relies more and more on the Internet, web applications are going to include very complex pattern recognition and data mining tasks that are required to be done online. Classification cascades are usually created manually using domain knowledge and are composed in most cases of two or three stages. In this chapter, a model-based algorithm of automatic generation of optimum classification cascades is devised. Given a large pool of classifiers (of size N), it builds a cascade that achieves the lowest possible recognition time while preserving the accuracy of the most powerful classifier in the pool. The proposed algorithm has a low complexity of O(N 2 ) where N is the number of classifiers in the pool. This gives us the freedom of using a large pool of classifiers which leads to more efficient cascades. Other cascade design techniques devised in the literature have very high complexity which hinders using large pool of classifiers.In this chapter we also analyze the performance of the devised algorithm showing its powerfulness and limitations. Also we present an algorithm for building a classification cascade of a given fixed length. This helps building cascades with space complexity constraints and helps in analyzing the performance of the devised algorithm for building optimum classification cascades.

show abstract

“…The selection of support vector machine models has enormously affected classification performance and computational complexity under different operational conditions. At present, most of the model selection was lack of good theoretical guidance [3][4]. The recent results in pattern recognition have shown that support vector machine (SVM) classifiers often have superior recognition rates in comparison to other classification methods.…”

Section: Introductionmentioning

confidence: 99%

Support Vector Machine for Mechanical Faults Diagnosis

Song

2010

2010 International Conference on Measuring Technology and Mechatronics Automation

View full text Add to dashboard Cite

Aiming at the difficulty that Support Vector Machine (SVM) model selection of classification algorithm affect classification accuracy, it research relevant factors that influence the precision of fault classifiers based on the typical fault data samples obtained by experimental setup of rotorbearing systems. The results show that different SVM classifiers, in which different kernel functions and different kernel functions parameters are adopted, will influence the precision of fault classifiers in conditions that fault data samples is small. It can be conveniently applied to choose appropriate kernel functions and kernel functions parameters in engineering application.

show abstract

An efficient three-stage classifier for handwritten digit recognition

Cited by 22 publications

References 6 publications

Isolated vietnamese handwriting recognition embedded system applied combined feature extraction method

Isolated vietnamese handwriting recognition embedded system applied combined feature extraction method

A Model-Based Approach for Building Optimum Classification Cascades

Support Vector Machine for Mechanical Faults Diagnosis

Contact Info

Product

Resources

About