| 1. IntroductionCharacter is very important to communication with other people. With the speeding up of the process of information, the application of handwritten character recognition needs to be more and more widely. But in the field of handwriting recognition ,it's still far from ideal, it should be strengthen handwritten character recognition technology .After long-term job, off-line handwritten character recognition research are more concerned . Such as the division of characters, pre-treatment technique have been made many achievements. However, in the identification of characters, off-line handwritten character recognition has a lot of trouble because the treatment of the characters are only two-dimensional lattice images, and also the reason of large deformation character , and so on, making up the difficult of handwritten character recognition .In addition, traditional statistical pattern recognition methods are more than enough at the number of samples carried out under the premise that only when the infinite number of samples to get good results. However, in practical problems, the sample size is usually limited, then the existing methods are difficult to achieve the desired results. Handwritten numeral recognition is divided into off-line identification and on-line recognition. One of the most complex is off-line handwritten character recognition. Mainly because of off-line handwritten character recognition process can not access to information when writting, so off-line handwritten numeral recognition is more challenging.Genetic Algorithm and Artificial Neural Network are parts of Intelligent Intormation Processing Technologies. Genetic Algorithm is high efficient, parallel, adaptive global optimization probability searching method. It provides an universal scheme which is solved complicated optimization and it produces strength robust. Artificial Neural Network has the ability of parallel processing, distributed infor-mation memory, self-adaptation, self-organization, associative learning and admitting-error. It can well solve the difficult of the recognition when the method of the writtien diffence. In a word, Intelligent Intormation Processing Technologies provide good method for uncertain, fuzzy, complicated information processing roblem such as characters recognition system.2. Research ContentIn the depth analysis and comparison of the current math symbols on the basis of identification methods, in order to set up a model of coarse grid characteristics, projection, cross-cut features, the structure of feature points as the characterize used niche genetic algorithms and neural networks to identify the mathematical symbols NGA-BP network . The main research contents are as follows:(1)A variety of digital methods of image pre-processing research, including images of the two values, filtering, thinning, normalized and so on, and the traditional method of image feature extraction and identification methods are discussed and presented in this article sign taken to identify the feature extraction method, that is based on statistical features and structural characteristics of the method of integration, extract the characteristics of coarse coarse grid characteristics, the projection features, cross-cut features, as well as structural characteristics of neural networks as input feature vectors.(2) Study the BP neural network of the principles of artificial neural models and artificial neural network topology and learning rules. Analysis of the BP network model and learning algorithm.(3) Research of the principles of Genetic Algorithms and Niche Genetic Algorithms. This dissertation studied the principles of Standard Genetic Algorithms and the process of manipulation. And detailedly introduces the genetic operations about selection, crossover and mutation. The principle and steps about Niche Genetic Algorithms (NGA) were deeply studied. And this paper compared the performance of Niche Genetic Algorithms with Standard Genetic Algorithms using Shubert test function. The result shows that the Niche Genetic Algorithms is valid.(4) According to the previous text, this paper presents a improved algorithm model based on NGA-BP which makes good use of searching virtue in overall range using genetic algorithm and great capability of searching in local range using error back-propagation algorithm to against the defects of BP neural network. according to the structure of BP neural network, the gene vectors are encoded with connection weights and thresholds of the neural networks, and ascertained the fitness function. Then, to search the global most optimization individual by using the operator such as roulette wheel selection and optimization saved and adaptive crossover and adaptive mutation and niche pass etc. When we find the most optimize individual, we decode it to connection weights and thresholds of the BP neural networks.(5)This paper presents a improved model based on niche genetic algorithm and BP artificial neural network to identify the mathematical symbols. The NGA-BP model which was established in the dissertation optimizes the weight and thresholds of the designed neural networks through the niche genetic algorithm. And then find the optimization initial connection weights and thresholds. It can improve the speed and stability of convergence of network. The simulation results of NGA-BP network model show that the speed of convergence and stability of convergence have greatly increased compared with BP network model. NGA-BP network model is better than the original BP network model improve on the convergence rate at 21.7%; Up on the success rate of convergence, NGA-BP network model improved by 16% than the BP network model;to the identification accuracy, NGA-BP network model of the recognition rate of 90.2%, significantly better than the BP network model of the recognition rate of 78.8%, taken note of this article NGA-BP network model to obtain good results.3. ConclusionBased on the deeply analysis of the traditional identification methods and research of artificial neural network and genetic algorithms which are used for identify of mathematical symbols, a algorithms which using genetic algorithm to optimize the network connecting value and threshold value of BP neural network is present in this paper. And a NGA-BP model for identification of mathematical symbols is established. Then train the NGA-BP network and fault diagnosis base on the founded model. The conclusions in the paper are as follows:1.The NGA-BP model which is established in this paper used for mathematical symbols identify,The model's recognition accuracy is 90.2%,which is satisfied the requirements of handwritten symbols identify. The simulation results show that the model is proper, feasible and correct for the identify of mathematical symbols.2. In this paper, the proposal method improved not only the speed of network convergence but also the stability of network convergence very well. The proposal method also improve the efficiency of fault and the precision of the identify of mathematical symbols. The algorithms presented in the paper can supply reference basis for the further research of mathematical symbols identify based on genetic algorithms and neural network and have a comparatively high generalization for other neural network research fields. |
PDF Full Text Request