Optimizing Recognition System of Persian Handwritten Digits

HOLOS

In this paper, Shuffled Frog Leaping Algorithm is used to improve the recognition rate of Persian handwritten digits. In proposed approach, the effective features in increasing the recognition rate are selected using the Binary Shuffled Frog Leaping Algorithm (BSFLA). By selecting the most suitable features from among all extracted features, the recognition rate is improved and computational costs are also decreased. The fitness function in BSFLA is the number of errors in the Fuzzy classifier which its minimum value is desired. The results indicate that Shuffled Frog Leaping algorithm (SFLA) is more efficient.

Handwritten character recognition systems using automated pattern recognition is one of the important issues in the field of Information Technology. In this paper, proposed a method based on combining artificial neural networks and genetic algorithms to recognize handwritten of the Persian and Arabic OFF-LINE characters. As the neural network searches for the optimal values of weights and biases of different layers the researcher used an intelligent genetic optimization algorithm to find optimal values. After preprocessing and feature extraction operation, BITMAP image as a file system entries from 10 different characters, each with 40 samples from Persian manuscript characters. A total of 400 different samples. 80% of samples are used for training (320 samples) and 20% of samples (80 samples) are used for network testing. Because there are lots of common ground between Arabic and Persian alphabet, Persian handwritten character recognition method is also applicable to the detection of Arabic words. The proposed method does not depend on a particular language and method, so it can be employed to recognize letters in different languages .It can also be used to identify letters' typed in a variety of languages. MSE obtained results of the combination of artificial neural networks and genetic algorithms showed that the proposed method is one of the best methods to employ in the field of pattern recognition.

Object recognition is an important research field of computer vision and has its application in a broad range of problems including image retrieval, compression, surveillance and medical diagnostics. The main goal of the object recognition problem is to recognize the objects of the same type even when they are viewed from different viewpoints. This goal, however, remains a challenge for computer vision to recognize objects having invariant features such as translation, rotation and scaling. Shape descriptors like Fourier and Moments are invariant with respect to transformation, rotation and scaling. Particle swarm optimization (PSO) is a population based soft computing technique. Particle Swarm Optimization technique shares numerous similarities with evolutionary computation techniques such as Genetic Algorithms (GA). One of the most important tasks regarding to object recognition is how to find number of descriptors of a given object. The query that arises is what is the optimum number of descriptors to be used with maximum recognition rate? , Are descriptors having equal importance? Such reasons signify the importance of these descriptors and also selecting the best descriptor by applying optimization technique. We have introduced an evolutionary optimization technique known as Genetic algorithm (GA) for solving the optimization problem. GA assigns, for each of these descriptors, a weighting factor that reflects the relative importance of that descriptor.

Image processing is an interesting and widely used field in the research area. Handwritten digit recognition is a sub module of the image processing. Although lots of research has been done so for in this field and still continues. This paper suggested a novel approach of off-line handwritten digit recognition. The paper classified the digits into four regions: the left part, right part, upper part and lower part. In these four parts the images are identified on the basis of curve. The curve of a left part image is converted into pixels and as well as rest of all parts. Then these pixels are compared through decision tree and the result of comparison describes the format of digit. This method is used to test the various handwritten digit form modified NIST and MNIST databases, which shows the great success rate.

2013

This work presents study and experimentation for object recognition when isolated objects are under discussion. The circumstances of similarity transformations, presence of noise, and occlusion have been included as the part of the study. For simplicity, instead of objects, outlines of the objects have been used for the whole process of the recognition. Fourier Descriptors have been used as features of the objects. From the analysis and results using Fourier Descriptors, the following questions arise: What is the optimum number of descriptors to be used? Are these descriptors of equal importance? To answer these questions, the problem of selecting the best descriptors has been formulated as an optimization problem. Particle Swarm Optimization technique has been mapped and used successfully to have an object recognition system using minimal number of Fourier Descriptors. The proposed method assigns, for each of these descriptors, a weighting factor that reflects the relative importance of that descriptor.

Ijca Proceedings on National Conference Cum Workshop on Bioinformatics and Computational Biology, 2014

The paper depicts the progress achieved in the field of character recognition using genetic algorithm. Character recognition is a process in image processing where the characters fed into the system are identified and classified. The main focus of this paper is on the offline character recognition since very less work has been done in this field. The use of genetic algorithm is the basis of this paper and it focuses on the advantages of using a genetic algorithm and also a survey of the works that have been implemented so far.

In this paper two approaches of genetic algorithm for feature subset selection are compared. The first approach considers a simple genetic algorithm (SGA) while the second one takes into account an iterative genetic algorithm (IGA) which is claimed to converge faster than SGA. Initially, we present an overview of the system to be optimized and the methodology applied in the experiments as well. Afterwards we discuss the advantages and drawbacks of each approach based on the experiments carried out on NIST SD19. Finally, we conclude that the IGA converges faster than the SGA, however, the SGA seems more suitable for our problem.

Applied Soft Computing, 2012

Identification of local regions from where optimal discriminating features can be extracted is one of the major tasks in the area of pattern recognition. To locate such regions different kind of region sampling techniques are used in the literature. There is no standard methodology to identify exactly such regions. Here we have proposed a methodology where local regions of varying heights and widths are created dynamically. Genetic algorithm (GA) is then applied on these local regions to sample the optimal set of local regions from where an optimal feature set can be extracted that has the best discriminating features. We have evaluated the proposed methodology on a data set of handwritten Bangla digits. In the present work, we have randomly generated seven sets of local regions and from every set, GA selects an optimal group of local regions which produces best recognition performance with a support vector machine (SVM) based classifier. Other popular optimization techniques like simulated annealing (SA) and hill climbing (HC) have also been evaluated with the same data set and maximum recognition accuracies were found to be 97%, 96.7% and 96.7% for GA, SA and HC, respectively. We have also compared the performance of the present technique with those of other zone based techniques on the same database.

Document binarization is an active research area for many years. Binarization algorithms play an important role in the preprocessing phase of any character recognition system. This paper compares several alternative binarization algorithms for handwritten documents, by evaluating their performance. The algorithms evaluated are, global thresholding, Otsu thresholding, Kittler-Illingworth and local thresholding, Niblack algorithm along with the proposed PSO algorithm. From the tests and results, we can wrap up with the assumption that the proposed algorithm shows improved results.

SYLWAN., 158(10)]. ISI Indexed, 2015

This manuscript considers a new approach based on two methods main swarm intelligence (SI), particle swarm optimization (PSO) and ant colony optimization (ACO) for solving recognition handwritten characters problem. We present an overview of the proposed approaches to be optimized and test on a number of handwritten characters in the experiments, as well. Experimental results show the higher degree performance of the proposed approaches. It is noted that a new approach in general generates an optimized comparison between the input samples and database samples which, improves the last recognition rate. Experimental results show that PSO algorithm outperforms ACO algorithm in convergence speed, coping with high dimensional spaces and straightforwardness with minimizing error recognition rate.