Khmer Character Recognition using Artificial Neural Network

Size: px
Start display at page:

Download "Khmer Character Recognition using Artificial Neural Network"

Transcription

1 Khmer Character Recognition using Artificial Neural Network Hann Meng * and Daniel Morariu * Faculty of Engineering, Lucian Blaga University of Sibiu, Sibiu, Romania meng.hann@rupp.edu.kh Tel: Advisor, Department of Computer Science and Electrical Engineering, Lucian Blaga University of Sibiu, Sibiu, Romania daniel.morariu@ulbsibiu.ro Abstract Character Recognition has become an interesting and a challenge topic research in the field of pattern recognition in recent decade. It has numerous applications including bank cheques, address sorting and conversion of handwritten or printed character into machine-readable form. Artificial neural network including self-organization map and multilayer perceptron network with the learning ability could offer the solution to character recognition problem. In this paper presents Khmer Character Recognition (KCR) system implemented in Matlab environment using artificial neural networks. The KCR system described the utilization of integrated self-organization map (SOM) network and multilayer perceptron (MLP) network with backpropagation learning algorithm for Khmer character recognition problem. Fig. 1 list of Khmer consonants I. INTRODUCTION Character recognition or optical character recognition (OCR) system is often used in machine translation, text-tospeech and text mining. It is been used to bring books, magazines, and printed documents to digital library by digitalizing them. The OCR application is the best example of pattern recognition. In order to be able to recognize Khmer scan text documents or handwritten documents, first important step is to be able to recognize Khmer characters. In this paper, I developed a Khmer character recognition system using artificial neural network. The system is implemented with self-organization map (SOM) network and multilayer feed-forward neural network using backpropagation-learning algorithm based on the scheme presented in Fig. 6. II. KHMER LANGUAGE CHARACTERISTICS Prepare y The Khmer script (Âksâr Khmêr) is used to write Khmer language (Cambodia official language). Cambodia officially known as the Kingdom of Cambodia that located in Southeast Asia. Ref. [12], Khmer differs from neighboring languages such as Thai, Lao, Burmese and Vietnamese. Khmer language has influenced by Sanskrit and Pali. A. Consonants Fig. 1 is a list of Khmer 33 consonants is shown below. Some characters are quite similar with little differences. Fig. 2. Numberical B. Numberical Khmer numerical are often used in literature more than handwritten which shown in Fig 2. In handwritten or every day used, people written as Arabic numerical which are widely known. III. BACKGROUND TECHNOLOGY In this section described the concept of artificial neural network (ANN). Artificial neural network has been use in many fields of study such as image processing, document clustering, segmentation, biology system, geography, chemistry, decision-making, data mining, optical character recognition (OCR), pattern recognition and many other interesting fields of research. Researchers have developed many artificial learning algorithms, which could enable the machine to have ability to perform and learn like human brain. ANN is trained and learned by example like human being and process information s like human brain. Ref. [4], The architecture of HNN is base on preclassification by using SOM (Self-Organization Map). Each preclassification is followed by the recognition modules, which PCA (Principle Component Analysis) cascaded with an MLP (Multilayer perceptron). A set of winner from SOM network is triggered the recognition modules to classify the input data. Voting mechanism is used to select winner when APSIPA APSIPA 2014

2 multiple winner occurred. By seeing the great potential of hierarchical neural network which integrate multiple artificial neural networks into character recognition system, selforganization map and multilayer perceptron neural network is chosen to implement our simple system. This section is organized as follows. Self-organizing map (SOM) is the clustering algorithm that maps input data to form grouping. Neural network is used as character classifier network and brief introduction of Matlab tool for artificial neural network. A. Self-Organization Map (SOM) Ref. [3, 9], the self-organizing map (SOM) network developed by Kohonen (1995), also called Kohonen neural network. Ref. [6], SOM is an effective network for analyzing of multidimensional data. SOM is one of Artificial Neural Network (ANN), which is suitable for data clustering. The SOM network consist of two layers of neuron: an input layer, where the inputs to the ANN are applied, and an output layer of neurons called competitive layer, where the grouping the inputs are formed. SOM is an unsupervised learning algorithm, which there is no desired output (no teacher), the network have to discovery itself the feature, pattern, clusters, regulation from the input data automatically. The idea this algorithm is to seek the winner unit from each input vector, and the modification for the synaptic coefficients is done for the winner unit and also for all units from the neighborhood of the winner unit. This network is like a map of an input dataset which shown in Fig 3. Algorithms 1). Initialize all weights with small random values (or initialize the weight using input data) 2). Set the initial learning rate and topological of neighborhood 3). Take the input sample (vector) and present to the network then calculate the distance between itself to all neurons in the network Calculate the distance of Best Matching Unit (BMU): x w!,! = min!,! x w!" (1) 4). The neuron that has the closest distance to the current sample is considered to be the winner neuron and update its weight. Update the weight: w! t + 1 = w! t + α t x w! t for i neigborhood w! otherwise 5). For all neuron found in the neighborhood of the winner neuron also update all their weights according to the Kohonen equation (2) 6). Repeat step 3-5 for all input vectors from the data set 7). Decrease the value of learning rate (α) and shrink the neighborhood (σ) (2) Fig. 3. The basic structure of self-organization map, ref. [6] 8). Repeat step 3 through 7 until the learning rate is less than a pre-specified threshold or a maximum number of iterations are reached. The proposed formula for computing of the Neighborhood: σ t = σ! e!!! And the Learning Rate: α t = α! e!!! Where t is the current age (stage), N is total number of iterations Note: The coefficient of neighborhood (σ! ) should be large enough because we want to start with a large number of neighborhoods at the beginning (recommended neighborhood be about 60-70% from the all neurons). The parameter of the learning rate (α! ) was chosen small value of 0.1 because we want that learning rate be small in order to descent quickly to zero. B. Neural Network Ref. [2], Warren McCulloch and Walter Pitts introduced a binary threshold as a computational model for neural network in Paul Werbos developed a learning algorithm called backpropagation of error in It is now used in many areas of studies and researches. Ref. [5], neural network is the most popular choice for developing character recognition system, which this network could learn well and provide high accuracy and speed for character identification. Neural network has been use to solve many character recognition problem; such as Chinese character recognition. A neural network contains a number of nodes (units or neurons) connected by edges. Each link has a weight associated with it. The weights can be a memory of the network and the learning process of the networks is to compute these weights so that the network presents best results suited to train data. Neural network is adaptive because it can adapt to the change of data and learn the characteristic of the input signal. The system is illustrated the noticeable and better result of using feed-forward network using back propagation algorithm with momentum and adaptive learning algorithm that provided better network classifier and higher accuracy in recognizing character.

3 Fig. 6. The structure of KCR System Fig. 4. Common model of artificial neural network x represent the inputs signal w represent the weights θ represent the bias value f is an activation function or cost function of the network that is known as sigmoid transfer function. The error output function is computation between actual output from the network and desired output. Error function is known as network s energy function. E =!! (d!!! o! )! (5) where d! is the desired output for neuron i th Fig. 5. Multilayer Perceptrons, ref. [1] Therefore, feed-forward neural network using backpropagation with adaptive learning is chosen to implement for Khmer character recognition system and used as character classification network. Computational Neural model The Fig. 4 show the common model of artificial neural network. Multilayer Perceptron Network and Backpropagation Multilayer perceptron is a feed-forward neural network. It consists of input layer, hidden layers and output layer. The common multilayer network has single hidden layer called two layer network in which count only number of hidden layers and output layer and input layer is not include. Multilayer perceptron network is the network that has more than one layer of perceptron with an activation function in each layer, which can learn any boolean function, Fig. 5. Compute the output in hidden and output layer in the following: Hidden layer h! = f(h!,where H! =! w!" x! + θ! (3) Output layer o! = f(o! ), where O! =! w!" h! + θ! (4) A. Application Flow Control IV. SYSTEM DESIGN The Fig. 6 shown the overall architecture of implemented Khmer character recognition system. The proposed Khmer character recognition (KCR) system scheme is adopted from ref. [4] that is depicted in figure above. The KCR system consists of two well-known artificial neural network methods that are integrated to solve the Khmer character recognition problem. Self-organization map is used for input data grouping and multilayer neural network is used for network classifier. The network is trained with set of character in 20 by 20 dimensional in black and white pixels. The input data images represent as 20 by 20 matrix. First is to transform input binary character image matrix to a vector with scalar values of 400 elements and then feed to the network. Then the SOM network has 400 neurons at the input layer, which is each neuron for each pixel. B. Data Representation Input Representation The input samples to the artificial neural network are mostly in image format. The input image can be from several sources such as scan text image, camera capture, handwritten or handprint document. As text image usually contains lot of

4 Fig. 7. Listed of 5 different fonts on a single character characters inside. Ref. [8, 10], to get the isolated character image from the text image, many methods are presented in articles and papers that are needed to do image preprocessing and segmentation. In our work, we do not adapt to segmentation and feature extraction methods above. The input image is isolated character image, which is predefined. Each input character image is an image with JPEG format, which is only need to convert to binary character image. Each character image is represented in 20x20 pixels that are 400 pixels. This is the feature selection for the network. The system is trained with 5 sets of different fonts which each font has 33 characters and 10 numerals. In total, there are 43 x 5 = 135 input samples. The Fig. 7 is shown 5 types of difference fonts on a single character in which each font has different styles, sizes and shapes that are used in different purposes. Some font provided a very thin of character or fat or bold and some are like handwriting. In the Fig. 8 is described the full process of input data transformation from character image to final data presentation for the networks. First, input character image is converted to binary matrix form in 20x20 dimensions, and then binary matrix is transformed to vector scalar values with 400 elements and put to dataset to represent all character matrixes. These processes is repeated until finish all the input character datasets. The transformation from character image to matrix is described in preprocessing section. Target Vector Encoding Multilayer Neural network is a supervised learning algorithm; in which in the training process is needed to know the desired output is needed. Supervise learning is called also learning with a teacher. When the actual output from the network is closer to the desired output, the small error is obtained. As the nature of neural network the output neurons are usually 0 or 1. There are several ways where target vector can represent in neural networks. In the case of Khmer language, there are 33 consonants, 25 dependent vowels, 14 independent vowels and 10 numerical. The total is = 82. So define the target vector with 82 dimensional is not a good choice. As a result, the method of defining the target vector above is hard to work on it and it is not a good solution to the problem. Therefore, we proposed target vector encoding technique to deal with this difficulty, see Fig. 9. Target vector encoding: 2! = numbers of possible target where n is number of digits. Fig. 8. Input data transformation Fig. 9. Target Vector Encoding Sample The idea behind the target vector encoding technique is to convert decimal index that represent of that character into binary form with specific number of digit. The n number of digit is requiring because it could show how many number of characters would be able to represent. For example, using only 5 digits, would have 2! = 2! = 32 possible targets to represent all 26 English alphabets. In Khmer language with 81 inputs by using 6 digits then the maximum number of target vector is 2! = 64. Using only 6 digits because in KCR system use only 43 (alphabets and numerals) inputs are cluster into 9 group base on it similarity by SOM network. So the using only 6 digits it could represent all input characters in each groups. More precisely see in the Table 3, which is shown the first 5 characters with 6 digits target vector encoding. This technique is flexible and scalable with less complexity for the neural network. C. Image Preprocessing Preprocessing is the stage to prepare data representation for putting to the network. In this section will describe about image to gray scale transformation and convert input character image to binary character matrix.

5 Fig. 10. Bitmap Image, a) isolate character, b) binary character image, c) character image matrix Gray Scale Transformation Gray scale transformation is a part of image processing. It converts color image into gray scale form. Ref. [11], usually images from the camera or scan documents are in the form of true color images, while most of the image processing techniques are worked with 256 gray scale image. Gray scale image also known as black and white and the value of color varying from weak (black) to strong intensity (white). Image Digitalizing Usually the documents are in the form printed document, handwritten character, typewriter and images in which human is able to visualize. In contrast to human, computer is worked on the machine encoding behind the image with a specific format. In artificial neural network, the input characters mostly are images, which are needed to convert into the form of binary character matrix, that is, the network able to learn. Image digitalizing is a process to transform images in the form of gray scale, color into binary form. In the Fig 10, the alphabet has been digitalizing into 20x20 = 400 digital cells, each cell have only a single color, either black or white. Each black pixel is assigned value to 0 and each white pixel is assigned value to 1, which is shown in Fig 10 (c). It is necessary to understand the information from the image to the matrix in a form meaningful to a machine encode information for computer. D. Additional System Features The additional features are use to enhance the system performance as well as the better recognition accuracy and speed. Data Normalization Normalization is the process to transform the input data for the network in the way that could highly system performance. Usually in artificial neural network the input data is normalized in the range of interval [0 1] or [-1 1]. Normalization can be made when create a new network by using minmax or decimal scaling function. For SOM network, it is better to transform the input value from 0 to -1 and 1 to 1. Since the normalization of input value would make the training faster and better data groups. For MLP neural network, the transformation input value from 0 to 0.1 and 1 to 0.9 for fast convergence and better precision since the sigmoid function is used in networks, which provide great achievement for network result. These transformations of the input data are based on experiences, since there is no specific Fig. 11. a) The original binary image, b) matrix represent character, and c) image with noise rule or algorithms to transform input data for a specific network. Adding Noise to the Data In reality, noise is happened unpredictable like in scan documents or handwritten documents. Noise is not a good thing, because it could limit the accuracy of network system. Our network is trained with the ideal data binary character images. But for testing, it is better to add some noises to the binary character images in order to know whether the network has ability to learn or not. Adding noise allows the system to explore and learn to approximate the best result. Thus adding noises to the binary character image is required for testing the network. Adding the noises randomly to the matrix I with the noise rate as following steps: 1) Calculate number of random points base on specific noise rate 2) Take random row index i, which i in range [1 20] 3) Take random column index j, which j in range [1 20] 4) Apply to the matrix I (i, j) with the random value in which value can be 0 or 1 5) Repeat step 2 to 4 until reach the number of random points The sample result by adding noises to original character image matrix is shown in Fig. 11 (c). E. Network Design The KCR system begins with the training of selforganization map network, which after training the data are clustered. Each cluster is trained with MLP neural network. Each cluster is trained corresponding to MLP neural network modules. The SOM network is trained using adaptive learning algorithm that is fast for data clustering and MLP network is trained using back propagation learning algorithm that is good for distinguishing character identity. Self-Organization Map Configuration SOM network consists of input layer and competitive layers and no bias. Ref. [7], authors used different size of SOM network (30x30 and 50x50) to test and find out the better recognition accuracy. Since in our system the dataset to train is very small, so SOM network size is used only 3x3, which input dataset group into 9 classes. Then each class is putted to each MLP network. It is used as preclassify the network. The network training parameters as following:

6 Input neurons: 400 (one sample) Output neurons: 9 clusters ([3 3]) Training algorithm: Random Order Weight/Bias Learning Rules Training epochs: 2000 Classifier of Neural Network Neural network is a powerful method for data modeling that could capture and represent the complex relationship between input and output. Neural network system architecture is resembled to perform task like human brain that is network acquired knowledge from learning (training) and stored it in the network s interconnected neurons, which is known as connection weights. Neural network is fast and realizable method for data classification toward achieving high recognition rate. Neural network with backpropagation algorithm consists of two hidden layers, input layer and output layer. The out come from each layer is activated with sigmoid transfer function. Output layer is the competitive layer, where the output character is identified. The initial weights, bias and momentum are initialized by the neural network system. The network training parameters as following: Input neurons: 400 (one sample) Hidden layer neuron: 100 for hidden layer 1, 50 for hidden layer 2 Output neurons: 6 (binary encoding) Training algorithm: gradient descent backpropagation with momentum and adaptive learning rate Performance function: Sum Squared Error Activation function: sigmoid transfer function Training goal achieved: 10e-29 Training epochs: 5000 Fig. 12. KCR Application Interface Training Interface The Fig 13 below is shown the training interface of Khmer character recognition system. The system is trained with predefine dataset in which consist of 5 set of fonts and each font contains 33 characters and 10 numerals. The system has two training phase: first training with SOM network and second training with MLP network. With this training interface user has possibility to set the training parameters of SOM in which consist of learning rate and training iterations and set the training parameters of MLP network in which consist of training goal, learning rate, training epochs, and number of hidden layer in both layer 1 and 2. The input dataset is first trained with self-organization map (SOM) in which data is group into subset. Each subset is fed to each corresponding multilayer perceptron and the training SOM network is saved after training is finished. The result of each grouping dataset from SOM is fed to the corresponding to train with MLP module and each trained MLP module is saved as recognition module. V. APPLICATION A. Application Interface The application interface of Khmer character recognition (KCR) system is implemented using Matlab software and is depicted in the Fig 12. The GUI (graphics user interface) has shown the four process from the input character to the predicte output character including several buttons where user can interact to test the system and configure to testing sitting. User be able to change the noise rate and number of character to be random for recognizing with the display text box for the recognition report Fig. 13. Training Interface

7 descent backpropagation with momentum and adaptive learning rate with the training goal 10e-29. Fig. 14. Testing Application Grouping dataset into subset is really important which can lead speed up training with MLP and better classification and high recognition accuracy. This is not only speed up the training but also speed up the testing process and provided high accuracy. Training with multilayer perceptron network (MLP) is the second phase and also train for network classifier. After each training module is finished, MLP network module is saved for network classifier that is used in the testing phase. B. Using Application The Khmer Character Recognition System (KCR) is implemented in Matlab, which is shown in the Fig. 14. By clinking on recognition button the character classification process to find the best match the input character corresponding to the noise rate is displayed on the predicted panel. The recognition module uses the number of testing randomly on an input character with the same noise rate. The system report text display shown the predicted character index and the number of frequency predicted character index found according to the number of random testing. The display predicted character is selected based on number of frequency index occurred. Each character has different forms, depending on the font it belongs. The results are display with only one standard font. The predicted result is displayed with the corresponding testing outcome index that match to the standard font index. The simulation with differences training goal 10e-20 and 10e-29 provided the same recognition rate when there is no noise added and when the noise added training goal with value 10e-29 provided higher recognition rate with all the different fonts in test set. Therefore, the training goal of neural network is used 10e-29 to train our system since the average recognition rate increase up to 94.13% with the same noise rate 10%. The summary of simulation approximate result of overall system that is tested with trained and untrained dataset from different fonts in which the network is trained using gradient VI. CONCLUSIONS AND FUTURE WORK The Khmer Character Recognition (KCR) system is integrated two artificial neural network techniques to work together namely self-organization map (SOM) and multilayer perceptron with backpropagation algorithm have been described in this paper. The preprocessing and preclassification (SOM) working together with character classifier modules could enhance the system performance, speed and recognition accuracy. As a result, using preclassification could reduce the complexity and allows faster recognizing the character since the result of preclassification is triggered a specific module of multilayer perceptron classifiers. That is the important feature in the system. The improvement of the system s performance by using backpropagation with momentum gradient descent learning term could retains quickly compare to the standard backpropagation in which using gradient descent learning term by experimented. The training goal is set to the maximum value (10e-29) since the average recognition rate increase from 72% (training goal 10e-10) to 94% recognition rate with the same noise rate. The average recognition of trained dataset resulted 65% of correct predictions and untrained dataset is approximately only 30% correct prediction with noise rate. The detail discussion of this paper could provide the beneficial conceptual of self-organization map and neural network as well as understanding the utilization and the implementation of the integrated of these two learning algorithms. This material could serve as a guide for readers working in the character recognition area. A lot of research is still needed to improve the current KCR system s performance with new feature. The KCR system is expected to be able to work with offline handwritten Khmer character, which the dataset is collected from difference writers. The scan images dataset would need to do image preprocessing (smoothing, noise remove, binarization), feature extraction and segmentation method to provide better recognition network system. By apply principle component analysis (PCA) that is a statistical method to projection the dataset from higher dimensions to lower dimension, which could enhance the system performance much faster. Therefore, the combine these methods with the existing integrated of self-organization map network and multilayer perceptron neural network may provide better solution to offline Khmer character recognition system.

8 REFERENCES [1] Brian D. Ripley, N. L. Hjort, Pattern Recognition and Neural Networks (1st ed.), Cambridge University Press, New York, NY, USA, 1995 [2] Gheorghita S., Munteanu, R., Enache M., Study of neural networks to improve performance for character recognition, AQTR IEEE International Conference on Automation Quality and Testing Robotics, pp.323, 326, May 2012 [3] Kohonen T., The self-organizing map, Proceedings of the IEEE, Vol.78, No.9, pp.1464, 1480, Sep 1990 [4] U. Halici, A. Erol, G. Ongun, Industrial Applications of Hierarchical Neural Networks: Character Recognition and Fingerprint Classification, CRC Press, Inc., Boca Raton, FL, USA, 1999 [5] Liangbin Zheng, Ruqi Chen, Xiaojin Cheng, Research on Offline Handwritten Chinese Character Recognition Based on BP Neural Networks, IPCSIT Vol. 51, IACSIT Press, Singapore, 2012 [6] M.H. Ghaseminezhad, A. Karami, A novel self-organizing map (SOM) neural network for discrete groups of data clustering, Applied Soft Computing, Vol 11, Issue 4, June 2011, Pages , ISSN , /j.asoc [7] Marinai S., Miotti B., Soda G., Bag of Characters and SOM Clustering for Script Recognition and Writer Identification, ICPR th International Conference on Pattern Recognition, pp.2182, 2185, Aug [8] Ranpreet Kaur, Baljit Singh, A Hybrid Neural Approach For Character Recognition System, International Journal of Computer Science and Information Technologies (IJCSIT), Vol. 2, , 2011 [9] Simon Haykin, Neural Networks: A Comprehensive Foundation (2nd ed.), Prentice Hall PTR, Upper Saddle River, NJ, USA, 1998 [10] Vijay Laxmi Sahu, Babita Kubde, Offline Handwritten Character Recognition Techniques using Neural Network, International Journal of Science and Research (IJSR), India Online ISSN: [11] Zhigang Zhang, Cong Wang, The Research of Vehicle Plate Recognition Technical Based on BP Neural Network, AASRI Procedia, Vol. 1, 2012, Pages 74-81, ISSN , /j.aasri [12] Khmer alphabet, Available from: Accessed: 1 May 2013

Cursive Handwriting Recognition System Using Feature Extraction and Artificial Neural Network

Cursive Handwriting Recognition System Using Feature Extraction and Artificial Neural Network Cursive Handwriting Recognition System Using Feature Extraction and Artificial Neural Network Utkarsh Dwivedi 1, Pranjal Rajput 2, Manish Kumar Sharma 3 1UG Scholar, Dept. of CSE, GCET, Greater Noida,

More information

International Journal of Advance Engineering and Research Development OPTICAL CHARACTER RECOGNITION USING ARTIFICIAL NEURAL NETWORK

International Journal of Advance Engineering and Research Development OPTICAL CHARACTER RECOGNITION USING ARTIFICIAL NEURAL NETWORK Scientific Journal of Impact Factor(SJIF): 4.14 International Journal of Advance Engineering and Research Development Special Issue for ICPCDECT 2016,Volume 3 Issue 1 e-issn(o): 2348-4470 p-issn(p): 2348-6406

More information

Mobile Application with Optical Character Recognition Using Neural Network

Mobile Application with Optical Character Recognition Using Neural Network Available Online at www.ijcsmc.com International Journal of Computer Science and Mobile Computing A Monthly Journal of Computer Science and Information Technology IJCSMC, Vol. 4, Issue. 1, January 2015,

More information

Neural Network Classifier for Isolated Character Recognition

Neural Network Classifier for Isolated Character Recognition Neural Network Classifier for Isolated Character Recognition 1 Ruby Mehta, 2 Ravneet Kaur 1 M.Tech (CSE), Guru Nanak Dev University, Amritsar (Punjab), India 2 M.Tech Scholar, Computer Science & Engineering

More information

HANDWRITTEN GURMUKHI CHARACTER RECOGNITION USING WAVELET TRANSFORMS

HANDWRITTEN GURMUKHI CHARACTER RECOGNITION USING WAVELET TRANSFORMS International Journal of Electronics, Communication & Instrumentation Engineering Research and Development (IJECIERD) ISSN 2249-684X Vol.2, Issue 3 Sep 2012 27-37 TJPRC Pvt. Ltd., HANDWRITTEN GURMUKHI

More information

Character Recognition Using Convolutional Neural Networks

Character Recognition Using Convolutional Neural Networks Character Recognition Using Convolutional Neural Networks David Bouchain Seminar Statistical Learning Theory University of Ulm, Germany Institute for Neural Information Processing Winter 2006/2007 Abstract

More information

An Algorithm For Training Multilayer Perceptron (MLP) For Image Reconstruction Using Neural Network Without Overfitting.

An Algorithm For Training Multilayer Perceptron (MLP) For Image Reconstruction Using Neural Network Without Overfitting. An Algorithm For Training Multilayer Perceptron (MLP) For Image Reconstruction Using Neural Network Without Overfitting. Mohammad Mahmudul Alam Mia, Shovasis Kumar Biswas, Monalisa Chowdhury Urmi, Abubakar

More information

Hand Written Character Recognition using VNP based Segmentation and Artificial Neural Network

Hand Written Character Recognition using VNP based Segmentation and Artificial Neural Network International Journal of Emerging Engineering Research and Technology Volume 4, Issue 6, June 2016, PP 38-46 ISSN 2349-4395 (Print) & ISSN 2349-4409 (Online) Hand Written Character Recognition using VNP

More information

Supervised Learning in Neural Networks (Part 2)

Supervised Learning in Neural Networks (Part 2) Supervised Learning in Neural Networks (Part 2) Multilayer neural networks (back-propagation training algorithm) The input signals are propagated in a forward direction on a layer-bylayer basis. Learning

More information

Traffic Signs Recognition using HP and HOG Descriptors Combined to MLP and SVM Classifiers

Traffic Signs Recognition using HP and HOG Descriptors Combined to MLP and SVM Classifiers Traffic Signs Recognition using HP and HOG Descriptors Combined to MLP and SVM Classifiers A. Salhi, B. Minaoui, M. Fakir, H. Chakib, H. Grimech Faculty of science and Technology Sultan Moulay Slimane

More information

CS6220: DATA MINING TECHNIQUES

CS6220: DATA MINING TECHNIQUES CS6220: DATA MINING TECHNIQUES Image Data: Classification via Neural Networks Instructor: Yizhou Sun yzsun@ccs.neu.edu November 19, 2015 Methods to Learn Classification Clustering Frequent Pattern Mining

More information

Handwritten Character Recognition with Feedback Neural Network

Handwritten Character Recognition with Feedback Neural Network Apash Roy et al / International Journal of Computer Science & Engineering Technology (IJCSET) Handwritten Character Recognition with Feedback Neural Network Apash Roy* 1, N R Manna* *Department of Computer

More information

SOMSN: An Effective Self Organizing Map for Clustering of Social Networks

SOMSN: An Effective Self Organizing Map for Clustering of Social Networks SOMSN: An Effective Self Organizing Map for Clustering of Social Networks Fatemeh Ghaemmaghami Research Scholar, CSE and IT Dept. Shiraz University, Shiraz, Iran Reza Manouchehri Sarhadi Research Scholar,

More information

Handwritten Gurumukhi Character Recognition by using Recurrent Neural Network

Handwritten Gurumukhi Character Recognition by using Recurrent Neural Network 139 Handwritten Gurumukhi Character Recognition by using Recurrent Neural Network Harmit Kaur 1, Simpel Rani 2 1 M. Tech. Research Scholar (Department of Computer Science & Engineering), Yadavindra College

More information

Handwritten Devanagari Character Recognition Model Using Neural Network

Handwritten Devanagari Character Recognition Model Using Neural Network Handwritten Devanagari Character Recognition Model Using Neural Network Gaurav Jaiswal M.Sc. (Computer Science) Department of Computer Science Banaras Hindu University, Varanasi. India gauravjais88@gmail.com

More information

Perceptrons and Backpropagation. Fabio Zachert Cognitive Modelling WiSe 2014/15

Perceptrons and Backpropagation. Fabio Zachert Cognitive Modelling WiSe 2014/15 Perceptrons and Backpropagation Fabio Zachert Cognitive Modelling WiSe 2014/15 Content History Mathematical View of Perceptrons Network Structures Gradient Descent Backpropagation (Single-Layer-, Multilayer-Networks)

More information

Image Compression: An Artificial Neural Network Approach

Image Compression: An Artificial Neural Network Approach Image Compression: An Artificial Neural Network Approach Anjana B 1, Mrs Shreeja R 2 1 Department of Computer Science and Engineering, Calicut University, Kuttippuram 2 Department of Computer Science and

More information

Multilayer Feed-forward networks

Multilayer Feed-forward networks Multi Feed-forward networks 1. Computational models of McCulloch and Pitts proposed a binary threshold unit as a computational model for artificial neuron. This first type of neuron has been generalized

More information

In this assignment, we investigated the use of neural networks for supervised classification

In this assignment, we investigated the use of neural networks for supervised classification Paul Couchman Fabien Imbault Ronan Tigreat Gorka Urchegui Tellechea Classification assignment (group 6) Image processing MSc Embedded Systems March 2003 Classification includes a broad range of decision-theoric

More information

Optical Character Recognition (OCR) for Printed Devnagari Script Using Artificial Neural Network

Optical Character Recognition (OCR) for Printed Devnagari Script Using Artificial Neural Network International Journal of Computer Science & Communication Vol. 1, No. 1, January-June 2010, pp. 91-95 Optical Character Recognition (OCR) for Printed Devnagari Script Using Artificial Neural Network Raghuraj

More information

Review on Methods of Selecting Number of Hidden Nodes in Artificial Neural Network

Review on Methods of Selecting Number of Hidden Nodes in Artificial Neural Network Available Online at www.ijcsmc.com International Journal of Computer Science and Mobile Computing A Monthly Journal of Computer Science and Information Technology IJCSMC, Vol. 3, Issue. 11, November 2014,

More information

COMPUTATIONAL INTELLIGENCE

COMPUTATIONAL INTELLIGENCE COMPUTATIONAL INTELLIGENCE Fundamentals Adrian Horzyk Preface Before we can proceed to discuss specific complex methods we have to introduce basic concepts, principles, and models of computational intelligence

More information

Classification and Regression using Linear Networks, Multilayer Perceptrons and Radial Basis Functions

Classification and Regression using Linear Networks, Multilayer Perceptrons and Radial Basis Functions ENEE 739Q SPRING 2002 COURSE ASSIGNMENT 2 REPORT 1 Classification and Regression using Linear Networks, Multilayer Perceptrons and Radial Basis Functions Vikas Chandrakant Raykar Abstract The aim of the

More information

Fingerprint Feature Extraction Based Discrete Cosine Transformation (DCT)

Fingerprint Feature Extraction Based Discrete Cosine Transformation (DCT) Fingerprint Feature Extraction Based Discrete Cosine Transformation (DCT) Abstract- Fingerprint identification and verification are one of the most significant and reliable identification methods. It is

More information

Recognition of Handwritten Digits using Machine Learning Techniques

Recognition of Handwritten Digits using Machine Learning Techniques Recognition of Handwritten Digits using Machine Learning Techniques Shobhit Srivastava #1, Sanjana Kalani #2,Umme Hani #3, Sayak Chakraborty #4 Department of Computer Science and Engineering Dayananda

More information

Simulation of Zhang Suen Algorithm using Feed- Forward Neural Networks

Simulation of Zhang Suen Algorithm using Feed- Forward Neural Networks Simulation of Zhang Suen Algorithm using Feed- Forward Neural Networks Ritika Luthra Research Scholar Chandigarh University Gulshan Goyal Associate Professor Chandigarh University ABSTRACT Image Skeletonization

More information

CHAPTER 8 COMPOUND CHARACTER RECOGNITION USING VARIOUS MODELS

CHAPTER 8 COMPOUND CHARACTER RECOGNITION USING VARIOUS MODELS CHAPTER 8 COMPOUND CHARACTER RECOGNITION USING VARIOUS MODELS 8.1 Introduction The recognition systems developed so far were for simple characters comprising of consonants and vowels. But there is one

More information

Simulation of Back Propagation Neural Network for Iris Flower Classification

Simulation of Back Propagation Neural Network for Iris Flower Classification American Journal of Engineering Research (AJER) e-issn: 2320-0847 p-issn : 2320-0936 Volume-6, Issue-1, pp-200-205 www.ajer.org Research Paper Open Access Simulation of Back Propagation Neural Network

More information

Artificial neural networks are the paradigm of connectionist systems (connectionism vs. symbolism)

Artificial neural networks are the paradigm of connectionist systems (connectionism vs. symbolism) Artificial Neural Networks Analogy to biological neural systems, the most robust learning systems we know. Attempt to: Understand natural biological systems through computational modeling. Model intelligent

More information

NOVATEUR PUBLICATIONS INTERNATIONAL JOURNAL OF INNOVATIONS IN ENGINEERING RESEARCH AND TECHNOLOGY [IJIERT] ISSN: VOLUME 5, ISSUE

NOVATEUR PUBLICATIONS INTERNATIONAL JOURNAL OF INNOVATIONS IN ENGINEERING RESEARCH AND TECHNOLOGY [IJIERT] ISSN: VOLUME 5, ISSUE OPTICAL HANDWRITTEN DEVNAGARI CHARACTER RECOGNITION USING ARTIFICIAL NEURAL NETWORK APPROACH JYOTI A.PATIL Ashokrao Mane Group of Institution, Vathar Tarf Vadgaon, India. DR. SANJAY R. PATIL Ashokrao Mane

More information

INTERNATIONAL JOURNAL OF PURE AND APPLIED RESEARCH IN ENGINEERING AND TECHNOLOGY

INTERNATIONAL JOURNAL OF PURE AND APPLIED RESEARCH IN ENGINEERING AND TECHNOLOGY INTERNATIONAL JOURNAL OF PURE AND APPLIED RESEARCH IN ENGINEERING AND TECHNOLOGY A PATH FOR HORIZING YOUR INNOVATIVE WORK HANDWRITTEN DEVANAGARI CHARACTERS RECOGNITION THROUGH SEGMENTATION AND ARTIFICIAL

More information

CHAPTER 1 INTRODUCTION

CHAPTER 1 INTRODUCTION CHAPTER 1 INTRODUCTION 1.1 Introduction Pattern recognition is a set of mathematical, statistical and heuristic techniques used in executing `man-like' tasks on computers. Pattern recognition plays an

More information

CLASSIFICATION WITH RADIAL BASIS AND PROBABILISTIC NEURAL NETWORKS

CLASSIFICATION WITH RADIAL BASIS AND PROBABILISTIC NEURAL NETWORKS CLASSIFICATION WITH RADIAL BASIS AND PROBABILISTIC NEURAL NETWORKS CHAPTER 4 CLASSIFICATION WITH RADIAL BASIS AND PROBABILISTIC NEURAL NETWORKS 4.1 Introduction Optical character recognition is one of

More information

OCR For Handwritten Marathi Script

OCR For Handwritten Marathi Script International Journal of Scientific & Engineering Research Volume 3, Issue 8, August-2012 1 OCR For Handwritten Marathi Script Mrs.Vinaya. S. Tapkir 1, Mrs.Sushma.D.Shelke 2 1 Maharashtra Academy Of Engineering,

More information

Text Extraction from Images

Text Extraction from Images International Journal of Electronic and Electrical Engineering. ISSN 0974-2174 Volume 7, Number 9 (2014), pp. 979-985 International Research Publication House http://www.irphouse.com Text Extraction from

More information

Artificial Neural Networks MLP, RBF & GMDH

Artificial Neural Networks MLP, RBF & GMDH Artificial Neural Networks MLP, RBF & GMDH Jan Drchal drchajan@fel.cvut.cz Computational Intelligence Group Department of Computer Science and Engineering Faculty of Electrical Engineering Czech Technical

More information

Segmentation of Kannada Handwritten Characters and Recognition Using Twelve Directional Feature Extraction Techniques

Segmentation of Kannada Handwritten Characters and Recognition Using Twelve Directional Feature Extraction Techniques Segmentation of Kannada Handwritten Characters and Recognition Using Twelve Directional Feature Extraction Techniques 1 Lohitha B.J, 2 Y.C Kiran 1 M.Tech. Student Dept. of ISE, Dayananda Sagar College

More information

Multi-Layer Perceptron Network For Handwritting English Character Recoginition

Multi-Layer Perceptron Network For Handwritting English Character Recoginition Multi-Layer Perceptron Network For Handwritting English Character Recoginition 1 Mohit Mittal, 2 Tarun Bhalla 1,2 Anand College of Engg & Mgmt., Kapurthala, Punjab, India Abstract Handwriting recognition

More information

Assignment # 5. Farrukh Jabeen Due Date: November 2, Neural Networks: Backpropation

Assignment # 5. Farrukh Jabeen Due Date: November 2, Neural Networks: Backpropation Farrukh Jabeen Due Date: November 2, 2009. Neural Networks: Backpropation Assignment # 5 The "Backpropagation" method is one of the most popular methods of "learning" by a neural network. Read the class

More information

A Survey of Problems of Overlapped Handwritten Characters in Recognition process for Gurmukhi Script

A Survey of Problems of Overlapped Handwritten Characters in Recognition process for Gurmukhi Script A Survey of Problems of Overlapped Handwritten Characters in Recognition process for Gurmukhi Script Arwinder Kaur 1, Ashok Kumar Bathla 2 1 M. Tech. Student, CE Dept., 2 Assistant Professor, CE Dept.,

More information

Two-step Modified SOM for Parallel Calculation

Two-step Modified SOM for Parallel Calculation Two-step Modified SOM for Parallel Calculation Two-step Modified SOM for Parallel Calculation Petr Gajdoš and Pavel Moravec Petr Gajdoš and Pavel Moravec Department of Computer Science, FEECS, VŠB Technical

More information

Learning. Learning agents Inductive learning. Neural Networks. Different Learning Scenarios Evaluation

Learning. Learning agents Inductive learning. Neural Networks. Different Learning Scenarios Evaluation Learning Learning agents Inductive learning Different Learning Scenarios Evaluation Slides based on Slides by Russell/Norvig, Ronald Williams, and Torsten Reil Material from Russell & Norvig, chapters

More information

Data Mining. Neural Networks

Data Mining. Neural Networks Data Mining Neural Networks Goals for this Unit Basic understanding of Neural Networks and how they work Ability to use Neural Networks to solve real problems Understand when neural networks may be most

More information

Hand Writing Numbers detection using Artificial Neural Networks

Hand Writing Numbers detection using Artificial Neural Networks Ahmad Saeed Mohammad 1 Dr. Ahmed Khalaf Hamoudi 2 Yasmin Abdul Ghani Abdul Kareem 1 1 Computer & Software Eng., College of Engineering, Al- Mustansiriya Univ., Baghdad, Iraq 2 Control & System Engineering,

More information

Neural network based Numerical digits Recognization using NNT in Matlab

Neural network based Numerical digits Recognization using NNT in Matlab Neural network based Numerical digits Recognization using NNT in Matlab ABSTRACT Amritpal kaur 1, Madhavi Arora 2 M.tech- ECE 1, Assistant Professor 2 Global institute of engineering and technology, Amritsar

More information

Pattern Classification Algorithms for Face Recognition

Pattern Classification Algorithms for Face Recognition Chapter 7 Pattern Classification Algorithms for Face Recognition 7.1 Introduction The best pattern recognizers in most instances are human beings. Yet we do not completely understand how the brain recognize

More information

Notes on Multilayer, Feedforward Neural Networks

Notes on Multilayer, Feedforward Neural Networks Notes on Multilayer, Feedforward Neural Networks CS425/528: Machine Learning Fall 2012 Prepared by: Lynne E. Parker [Material in these notes was gleaned from various sources, including E. Alpaydin s book

More information

OMBP: Optic Modified BackPropagation training algorithm for fast convergence of Feedforward Neural Network

OMBP: Optic Modified BackPropagation training algorithm for fast convergence of Feedforward Neural Network 2011 International Conference on Telecommunication Technology and Applications Proc.of CSIT vol.5 (2011) (2011) IACSIT Press, Singapore OMBP: Optic Modified BackPropagation training algorithm for fast

More information

Handwritten character and word recognition using their geometrical features through neural networks

Handwritten character and word recognition using their geometrical features through neural networks Handwritten character and word recognition using their geometrical features through neural networks Sudarshan Sawant 1, Prof. Seema Baji 2 1 Student, Department of electronics and Tele-communications,

More information

HCR Using K-Means Clustering Algorithm

HCR Using K-Means Clustering Algorithm HCR Using K-Means Clustering Algorithm Meha Mathur 1, Anil Saroliya 2 Amity School of Engineering & Technology Amity University Rajasthan, India Abstract: Hindi is a national language of India, there are

More information

Visual object classification by sparse convolutional neural networks

Visual object classification by sparse convolutional neural networks Visual object classification by sparse convolutional neural networks Alexander Gepperth 1 1- Ruhr-Universität Bochum - Institute for Neural Dynamics Universitätsstraße 150, 44801 Bochum - Germany Abstract.

More information

Radial Basis Function Neural Network Classifier

Radial Basis Function Neural Network Classifier Recognition of Unconstrained Handwritten Numerals by a Radial Basis Function Neural Network Classifier Hwang, Young-Sup and Bang, Sung-Yang Department of Computer Science & Engineering Pohang University

More information

6. NEURAL NETWORK BASED PATH PLANNING ALGORITHM 6.1 INTRODUCTION

6. NEURAL NETWORK BASED PATH PLANNING ALGORITHM 6.1 INTRODUCTION 6 NEURAL NETWORK BASED PATH PLANNING ALGORITHM 61 INTRODUCTION In previous chapters path planning algorithms such as trigonometry based path planning algorithm and direction based path planning algorithm

More information

Asst. Prof. Bhagwat Kakde

Asst. Prof. Bhagwat Kakde Volume 3, Issue 11, November 2013 ISSN: 2277 128X International Journal of Advanced Research in Computer Science and Software Engineering Research Paper Available online at: www.ijarcsse.com A Novel Approach

More information

Artificial Intelligence Introduction Handwriting Recognition Kadir Eren Unal ( ), Jakob Heyder ( )

Artificial Intelligence Introduction Handwriting Recognition Kadir Eren Unal ( ), Jakob Heyder ( ) Structure: 1. Introduction 2. Problem 3. Neural network approach a. Architecture b. Phases of CNN c. Results 4. HTM approach a. Architecture b. Setup c. Results 5. Conclusion 1.) Introduction Artificial

More information

Classification of Printed Chinese Characters by Using Neural Network

Classification of Printed Chinese Characters by Using Neural Network Classification of Printed Chinese Characters by Using Neural Network ATTAULLAH KHAWAJA Ph.D. Student, Department of Electronics engineering, Beijing Institute of Technology, 100081 Beijing, P.R.CHINA ABDUL

More information

SEVERAL METHODS OF FEATURE EXTRACTION TO HELP IN OPTICAL CHARACTER RECOGNITION

SEVERAL METHODS OF FEATURE EXTRACTION TO HELP IN OPTICAL CHARACTER RECOGNITION SEVERAL METHODS OF FEATURE EXTRACTION TO HELP IN OPTICAL CHARACTER RECOGNITION Binod Kumar Prasad * * Bengal College of Engineering and Technology, Durgapur, W.B., India. Rajdeep Kundu 2 2 Bengal College

More information

Argha Roy* Dept. of CSE Netaji Subhash Engg. College West Bengal, India.

Argha Roy* Dept. of CSE Netaji Subhash Engg. College West Bengal, India. Volume 3, Issue 3, March 2013 ISSN: 2277 128X International Journal of Advanced Research in Computer Science and Software Engineering Research Paper Available online at: www.ijarcsse.com Training Artificial

More information

A Technique for Offline Handwritten Character Recognition

A Technique for Offline Handwritten Character Recognition A Technique for Offline Handwritten Character Recognition 1 Shilpy Bansal, 2 Mamta Garg, 3 Munish Kumar 1 Lecturer, Department of Computer Science Engineering, BMSCET, Muktsar, Punjab 2 Assistant Professor,

More information

A System for Joining and Recognition of Broken Bangla Numerals for Indian Postal Automation

A System for Joining and Recognition of Broken Bangla Numerals for Indian Postal Automation A System for Joining and Recognition of Broken Bangla Numerals for Indian Postal Automation K. Roy, U. Pal and B. B. Chaudhuri CVPR Unit; Indian Statistical Institute, Kolkata-108; India umapada@isical.ac.in

More information

Fingerprint Identification System Based On Neural Network

Fingerprint Identification System Based On Neural Network Fingerprint Identification System Based On Neural Network Mr. Lokhande S.K., Prof. Mrs. Dhongde V.S. ME (VLSI & Embedded Systems), Vishwabharati Academy s College of Engineering, Ahmednagar (MS), India

More information

Rough Set Approach to Unsupervised Neural Network based Pattern Classifier

Rough Set Approach to Unsupervised Neural Network based Pattern Classifier Rough Set Approach to Unsupervised Neural based Pattern Classifier Ashwin Kothari, Member IAENG, Avinash Keskar, Shreesha Srinath, and Rakesh Chalsani Abstract Early Convergence, input feature space with

More information

Neural Networks. CE-725: Statistical Pattern Recognition Sharif University of Technology Spring Soleymani

Neural Networks. CE-725: Statistical Pattern Recognition Sharif University of Technology Spring Soleymani Neural Networks CE-725: Statistical Pattern Recognition Sharif University of Technology Spring 2013 Soleymani Outline Biological and artificial neural networks Feed-forward neural networks Single layer

More information

Cluster analysis of 3D seismic data for oil and gas exploration

Cluster analysis of 3D seismic data for oil and gas exploration Data Mining VII: Data, Text and Web Mining and their Business Applications 63 Cluster analysis of 3D seismic data for oil and gas exploration D. R. S. Moraes, R. P. Espíndola, A. G. Evsukoff & N. F. F.

More information

Machine Learning 13. week

Machine Learning 13. week Machine Learning 13. week Deep Learning Convolutional Neural Network Recurrent Neural Network 1 Why Deep Learning is so Popular? 1. Increase in the amount of data Thanks to the Internet, huge amount of

More information

Handwritten Hindi Character Recognition System Using Edge detection & Neural Network

Handwritten Hindi Character Recognition System Using Edge detection & Neural Network Handwritten Hindi Character Recognition System Using Edge detection & Neural Network Tanuja K *, Usha Kumari V and Sushma T M Acharya Institute of Technology, Bangalore, India Abstract Handwritten recognition

More information

Integrating Text Mining with Image Processing

Integrating Text Mining with Image Processing IOSR Journal of Computer Engineering (IOSR-JCE) e-issn: 2278-0661,p-ISSN: 2278-8727 PP 01-05 www.iosrjournals.org Integrating Text Mining with Image Processing Anjali Sahu 1, Pradnya Chavan 2, Dr. Suhasini

More information

Iris recognition using SVM and BP algorithms

Iris recognition using SVM and BP algorithms International Journal of Engineering Research and Advanced Technology (IJERAT) DOI: http://dx.doi.org/10.31695/ijerat.2018.3262 E-ISSN : 2454-6135 Volume.4, Issue 5 May -2018 Iris recognition using SVM

More information

Figure (5) Kohonen Self-Organized Map

Figure (5) Kohonen Self-Organized Map 2- KOHONEN SELF-ORGANIZING MAPS (SOM) - The self-organizing neural networks assume a topological structure among the cluster units. - There are m cluster units, arranged in a one- or two-dimensional array;

More information

Yuki Osada Andrew Cannon

Yuki Osada Andrew Cannon Yuki Osada Andrew Cannon 1 Humans are an intelligent species One feature is the ability to learn The ability to learn comes down to the brain The brain learns from experience Research shows that the brain

More information

A Neural Network Model Of Insurance Customer Ratings

A Neural Network Model Of Insurance Customer Ratings A Neural Network Model Of Insurance Customer Ratings Jan Jantzen 1 Abstract Given a set of data on customers the engineering problem in this study is to model the data and classify customers

More information

LECTURE NOTES Professor Anita Wasilewska NEURAL NETWORKS

LECTURE NOTES Professor Anita Wasilewska NEURAL NETWORKS LECTURE NOTES Professor Anita Wasilewska NEURAL NETWORKS Neural Networks Classifier Introduction INPUT: classification data, i.e. it contains an classification (class) attribute. WE also say that the class

More information

An Intelligent Technique for Image Compression

An Intelligent Technique for Image Compression An Intelligent Technique for Image Compression Athira Mayadevi Somanathan 1, V. Kalaichelvi 2 1 Dept. Of Electronics and Communications Engineering, BITS Pilani, Dubai, U.A.E. 2 Dept. Of Electronics and

More information

Assignment 2. Classification and Regression using Linear Networks, Multilayer Perceptron Networks, and Radial Basis Functions

Assignment 2. Classification and Regression using Linear Networks, Multilayer Perceptron Networks, and Radial Basis Functions ENEE 739Q: STATISTICAL AND NEURAL PATTERN RECOGNITION Spring 2002 Assignment 2 Classification and Regression using Linear Networks, Multilayer Perceptron Networks, and Radial Basis Functions Aravind Sundaresan

More information

A Multimodal Framework for the Recognition of Ancient Tamil Handwritten Characters in Palm Manuscript Using Boolean Bitmap Pattern of Image Zoning

A Multimodal Framework for the Recognition of Ancient Tamil Handwritten Characters in Palm Manuscript Using Boolean Bitmap Pattern of Image Zoning A Multimodal Framework for the Recognition of Ancient Tamil Handwritten s in Palm Manuscript Using Boolean Bitmap Pattern of Zoning E.K.Vellingiriraj, Asst. Professor and Dr.P.Balasubramanie, Professor

More information

Climate Precipitation Prediction by Neural Network

Climate Precipitation Prediction by Neural Network Journal of Mathematics and System Science 5 (205) 207-23 doi: 0.7265/259-529/205.05.005 D DAVID PUBLISHING Juliana Aparecida Anochi, Haroldo Fraga de Campos Velho 2. Applied Computing Graduate Program,

More information

ECG782: Multidimensional Digital Signal Processing

ECG782: Multidimensional Digital Signal Processing ECG782: Multidimensional Digital Signal Processing Object Recognition http://www.ee.unlv.edu/~b1morris/ecg782/ 2 Outline Knowledge Representation Statistical Pattern Recognition Neural Networks Boosting

More information

CMPT 882 Week 3 Summary

CMPT 882 Week 3 Summary CMPT 882 Week 3 Summary! Artificial Neural Networks (ANNs) are networks of interconnected simple units that are based on a greatly simplified model of the brain. ANNs are useful learning tools by being

More information

Recognition of Gurmukhi Text from Sign Board Images Captured from Mobile Camera

Recognition of Gurmukhi Text from Sign Board Images Captured from Mobile Camera International Journal of Information & Computation Technology. ISSN 0974-2239 Volume 4, Number 17 (2014), pp. 1839-1845 International Research Publications House http://www. irphouse.com Recognition of

More information

Handwritten Hindi Numerals Recognition System

Handwritten Hindi Numerals Recognition System CS365 Project Report Handwritten Hindi Numerals Recognition System Submitted by: Akarshan Sarkar Kritika Singh Project Mentor: Prof. Amitabha Mukerjee 1 Abstract In this project, we consider the problem

More information

NEURAL NETWORK-BASED SEGMENTATION OF TEXTURES USING GABOR FEATURES

NEURAL NETWORK-BASED SEGMENTATION OF TEXTURES USING GABOR FEATURES NEURAL NETWORK-BASED SEGMENTATION OF TEXTURES USING GABOR FEATURES A. G. Ramakrishnan, S. Kumar Raja, and H. V. Raghu Ram Dept. of Electrical Engg., Indian Institute of Science Bangalore - 560 012, India

More information

An Efficient Character Segmentation Based on VNP Algorithm

An Efficient Character Segmentation Based on VNP Algorithm Research Journal of Applied Sciences, Engineering and Technology 4(24): 5438-5442, 2012 ISSN: 2040-7467 Maxwell Scientific organization, 2012 Submitted: March 18, 2012 Accepted: April 14, 2012 Published:

More information

Character Recognition of High Security Number Plates Using Morphological Operator

Character Recognition of High Security Number Plates Using Morphological Operator Character Recognition of High Security Number Plates Using Morphological Operator Kamaljit Kaur * Department of Computer Engineering, Baba Banda Singh Bahadur Polytechnic College Fatehgarh Sahib,Punjab,India

More information

Classification Lecture Notes cse352. Neural Networks. Professor Anita Wasilewska

Classification Lecture Notes cse352. Neural Networks. Professor Anita Wasilewska Classification Lecture Notes cse352 Neural Networks Professor Anita Wasilewska Neural Networks Classification Introduction INPUT: classification data, i.e. it contains an classification (class) attribute

More information

A Review on Handwritten Character Recognition

A Review on Handwritten Character Recognition IJCST Vo l. 8, Is s u e 1, Ja n - Ma r c h 2017 ISSN : 0976-8491 (Online) ISSN : 2229-4333 (Print) A Review on Handwritten Character Recognition 1 Anisha Sharma, 2 Soumil Khare, 3 Sachin Chavan 1,2,3 Dept.

More information

Supervised Learning (contd) Linear Separation. Mausam (based on slides by UW-AI faculty)

Supervised Learning (contd) Linear Separation. Mausam (based on slides by UW-AI faculty) Supervised Learning (contd) Linear Separation Mausam (based on slides by UW-AI faculty) Images as Vectors Binary handwritten characters Treat an image as a highdimensional vector (e.g., by reading pixel

More information

Isolated Curved Gurmukhi Character Recognition Using Projection of Gradient

Isolated Curved Gurmukhi Character Recognition Using Projection of Gradient International Journal of Computational Intelligence Research ISSN 0973-1873 Volume 13, Number 6 (2017), pp. 1387-1396 Research India Publications http://www.ripublication.com Isolated Curved Gurmukhi Character

More information

Channel Performance Improvement through FF and RBF Neural Network based Equalization

Channel Performance Improvement through FF and RBF Neural Network based Equalization Channel Performance Improvement through FF and RBF Neural Network based Equalization Manish Mahajan 1, Deepak Pancholi 2, A.C. Tiwari 3 Research Scholar 1, Asst. Professor 2, Professor 3 Lakshmi Narain

More information

International Journal of Advance Research in Engineering, Science & Technology

International Journal of Advance Research in Engineering, Science & Technology Impact Factor (SJIF): 4.542 International Journal of Advance Research in Engineering, Science & Technology e-issn: 2393-9877, p-issn: 2394-2444 Volume 4, Issue 4, April-2017 A Simple Effective Algorithm

More information

A Data Classification Algorithm of Internet of Things Based on Neural Network

A Data Classification Algorithm of Internet of Things Based on Neural Network A Data Classification Algorithm of Internet of Things Based on Neural Network https://doi.org/10.3991/ijoe.v13i09.7587 Zhenjun Li Hunan Radio and TV University, Hunan, China 278060389@qq.com Abstract To

More information

Extract an Essential Skeleton of a Character as a Graph from a Character Image

Extract an Essential Skeleton of a Character as a Graph from a Character Image Extract an Essential Skeleton of a Character as a Graph from a Character Image Kazuhisa Fujita University of Electro-Communications 1-5-1 Chofugaoka, Chofu, Tokyo, 182-8585 Japan k-z@nerve.pc.uec.ac.jp

More information

Unsupervised Learning

Unsupervised Learning Networks for Pattern Recognition, 2014 Networks for Single Linkage K-Means Soft DBSCAN PCA Networks for Kohonen Maps Linear Vector Quantization Networks for Problems/Approaches in Machine Learning Supervised

More information

A Document Image Analysis System on Parallel Processors

A Document Image Analysis System on Parallel Processors A Document Image Analysis System on Parallel Processors Shamik Sural, CMC Ltd. 28 Camac Street, Calcutta 700 016, India. P.K.Das, Dept. of CSE. Jadavpur University, Calcutta 700 032, India. Abstract This

More information

Anale. Seria Informatică. Vol. XVII fasc Annals. Computer Science Series. 17 th Tome 1 st Fasc. 2019

Anale. Seria Informatică. Vol. XVII fasc Annals. Computer Science Series. 17 th Tome 1 st Fasc. 2019 EVALUATION OF AN OPTICAL CHARACTER RECOGNITION MODEL FOR YORUBA TEXT 1 Abimbola Akintola, 2 Tunji Ibiyemi, 3 Amos Bajeh 1,3 Department of Computer Science, University of Ilorin, Nigeria 2 Department of

More information

Linear Discriminant Analysis in Ottoman Alphabet Character Recognition

Linear Discriminant Analysis in Ottoman Alphabet Character Recognition Linear Discriminant Analysis in Ottoman Alphabet Character Recognition ZEYNEB KURT, H. IREM TURKMEN, M. ELIF KARSLIGIL Department of Computer Engineering, Yildiz Technical University, 34349 Besiktas /

More information

Lecture 2 Notes. Outline. Neural Networks. The Big Idea. Architecture. Instructors: Parth Shah, Riju Pahwa

Lecture 2 Notes. Outline. Neural Networks. The Big Idea. Architecture. Instructors: Parth Shah, Riju Pahwa Instructors: Parth Shah, Riju Pahwa Lecture 2 Notes Outline 1. Neural Networks The Big Idea Architecture SGD and Backpropagation 2. Convolutional Neural Networks Intuition Architecture 3. Recurrent Neural

More information

INTERNATIONAL RESEARCH JOURNAL OF MULTIDISCIPLINARY STUDIES

INTERNATIONAL RESEARCH JOURNAL OF MULTIDISCIPLINARY STUDIES STUDIES & SPPP's, Karmayogi Engineering College, Pandharpur Organize National Conference Special Issue March 2016 Neuro-Fuzzy System based Handwritten Marathi System Numerals Recognition 1 Jayashri H Patil(Madane),

More information

Scene Text Detection Using Machine Learning Classifiers

Scene Text Detection Using Machine Learning Classifiers 601 Scene Text Detection Using Machine Learning Classifiers Nafla C.N. 1, Sneha K. 2, Divya K.P. 3 1 (Department of CSE, RCET, Akkikkvu, Thrissur) 2 (Department of CSE, RCET, Akkikkvu, Thrissur) 3 (Department

More information

FEATURE EXTRACTION TECHNIQUE FOR HANDWRITTEN INDIAN NUMBERS CLASSIFICATION

FEATURE EXTRACTION TECHNIQUE FOR HANDWRITTEN INDIAN NUMBERS CLASSIFICATION FEATURE EXTRACTION TECHNIQUE FOR HANDWRITTEN INDIAN NUMBERS CLASSIFICATION 1 SALAMEH A. MJLAE, 2 SALIM A. ALKHAWALDEH, 3 SALAH M. AL-SALEH 1, 3 Department of Computer Science, Zarqa University Collage,

More information

Devanagari Handwriting Recognition and Editing Using Neural Network

Devanagari Handwriting Recognition and Editing Using Neural Network Devanagari Handwriting Recognition and Editing Using Neural Network Sohan Lal Sahu RSR Rungta College of Engineering & Technology (RSR-RCET), Bhilai 490024 Abstract- Character recognition plays an important

More information