1,557 Hits in 2.3 sec

A Study on Japanese Historical Character Recognition Using Modular Neural Networks

Tadashi Horiuchi, Satoru Kato
2009 2009 Fourth International Conference on Innovative Computing, Information and Control (ICICIC)  
The modular neural networks consist of two kinds of classifiers: a rough-classifier and a set of fine-classifiers.  ...  The final result of character recognition is derived by selecting the MLP which has the maximum output among the set of MLPs.  ...  Modular Neural Networks for Historical Character Recognition.  ... 
doi:10.1109/icicic.2009.57 fatcat:gwsamuhmk5ejhiykmuhmgabo6y

Evaluating a Zoning Mechanism and Class-Modular Architecture for Handwritten Characters Recognition [chapter]

Sandra de Avila, Leonardo Matos, Cinthia Freitas, João M. de Carvalho
2007 Lecture Notes in Computer Science  
Other set of experiments utilized the IRONOFF database resulting in recognition rates of 89.21% and 80.75% for uppercase and lowercase characters respectively, also with the class-modular MLP.  ...  Two architectures were tested and evaluated: a conventional MLP (Multiple Layer Perceptron) and a class-modular MLP.  ...  Multi-Layer Perceptron Neural networks (MLP-NN) classifiers have proven to be powerful tools in pattern recognition [14] .  ... 
doi:10.1007/978-3-540-76725-1_54 fatcat:gfifmuzvsbaqtbscmhixjhqmgy

Dynamic Zoning Selection for Handwritten Character Recognition [chapter]

Luciane Y. Hirabara, Simone B. K. Aires, Cinthia O. A. Freitas, Alceu S. Britto, Robert Sabourin
2011 Lecture Notes in Computer Science  
The information provided by the first level drives the second level in the selection of the appropriate feature extraction method and the corresponding class-modular neural network.  ...  The experimental protocol has shown significant recognition rates for handwritten characters (from 80.82% to 88.13%).  ...  -CR) Table 1 . 1 Conventional NN and Class-Modular MLP-NN Characters Conventional NN Recognition Rate (%) Class-Modular MLP-NN Recognition Rate (%) ZGlobal Z4 Z5H Z5V Z7 Z4 Z5H Z5V Z7  ... 
doi:10.1007/978-3-642-25085-9_60 fatcat:7ygukuq2pvfirawlt5dgrgdoya

The Multi-Class Imbalance Problem: Cost Functions with Modular and Non-Modular Neural Networks [chapter]

Roberto Alejo, Jose M. Sotoca, R. M. Valdovinos, Gustavo A. Casañ
2009 Advances in Soft Computing  
In this paper, the behavior of Modular and Non-Modular Neural Networks trained with the classical backpropagation algorithm in batch mode and applied to classification problems with Multi-Class imbalance  ...  So, in our Mod-NN, given a instance test x i , the two class network with the highest rating is taken as the class label for that instance.  ...  Multilayer Perceptron (MLP). MLP and RBFNN are two well-known NN in the pattern recognition field [10] .  ... 
doi:10.1007/978-3-642-01216-7_44 dblp:conf/isnn/AlejoSVC09 fatcat:wfjuoxjnbzaezmjdz2vneujgsy

Combining linear discriminant functions with neural networks for supervised learning

Ke Chen, Xiang Yu, Huisheng Chi
1997 Neural computing & applications (Print)  
The proposed architecture provides an efficient way to apply existing neural networks (e.g. multi-layered perceptron) for solving a large scale problem.  ...  For constructive learning, growing and credit-assignment algorithms are developed to serve for the hybrid architecture.  ...  We wish to thank Liping Yang for valuable and constructive discussions, as well as for providing a program on the Levenberg-Marquat algorithm for simulation.  ... 
doi:10.1007/bf01670150 fatcat:s6spr4y7dvgdjil23evjrwwanu


2012 International Journal of Machine Intelligence  
Results show that this new neural network model is more accurate than the other NN models. These results suggest that this model is effective for classification of satellite image data.  ...  This paper describes the MLP NN classifier performing optimally in classifying the different land types from Landsat data.  ...  These networks process their input using several parallel MLPs and then recombine the results. In contrast to the MLP, modular networks do not have full interconnectivity between their layers.  ... 
doi:10.9735/0975-2927.4.2.414-420 fatcat:ksc2uoapm5fjzawsm4wlw6lnje

From data topology to a modular classifier

Abdellatif Ennaji, Arnaud Ribert, Yves Lecourtier
2003 International Journal on Document Analysis and Recognition  
Experimental results for the handwritten digit recognition problem and comparison with neural and statistical nonmodular classifiers are given.  ...  This article describes an approach to designing a distributed and modular neural classifier.  ...  Neural networks, and more particularly Multi-Layer Perceptrons (MLP) [8, 15, 33, 34] have received a great deal of attention.  ... 
doi:10.1007/s10032-002-0095-3 fatcat:ipq45qmnhzbi5cb764tppce4nu

Usage of neural networks in ubiquitous computing systems

Andrey V. Gavrilov
2008 2008 Third International Forum on Strategic Technologies  
One of most perspective techniques for sensing in ubiquitous computing systems is neural networks.  ...  In this paper we describe features of usage of neural networks in ubiquitous computing and its implementation for solving of some tasks in middleware ubiquitous computing system for smart environment.  ...  In our modular neural network we use some multi-layer perceptrons (MLP) for different subsets of visible access points, stored in so-called matrix of visibility obtained during site calibration.  ... 
doi:10.1109/ifost.2008.4603019 fatcat:n7sd3wz5wbbjrkkmvouacajwpa


Arif Iqbal Mozumder
2017 International Journal of Advanced Research in Computer Science  
This paper proposes an approach for recognition of an iris using Modular Neural Network and Fuzzy inference system based Score Level Fusion.  ...  Image Acquisition, Feature Extraction, Segmentation and Recognition.  ...  architecture of the module used in the present work consists of five small Multi-Layer Perceptron (MLP) neural networks.  ... 
doi:10.26483/ijarcs.v8i7.4260 fatcat:ihko5ofkfjbz3fwkce67l3alpi

A modular classification model for received signal strength based location systems

Uzair Ahmad, Andrey V. Gavrilov, Sungyoung Lee, Young-Koo Lee
2008 Neurocomputing  
Previously we reported small scale prototype location estimation system based on modular multi layer perceptron (MLP) networks [2, 1] .  ...  Fig. 12 shows an arbitrary structure of an MLP network for location estimation.  ... 
doi:10.1016/j.neucom.2007.11.045 fatcat:ka47dxu7rjfblp7baaq2voc52e

A review on various techniques used to recognize off-line handwritten Malayalam characters

M.P. Ayyoob, P. Muhamed Ilyas
2021 Malaya Journal of Matematik  
This paper presents a comprehensive survey of various techniques with its accuracy rate on Malayalam off-line handwriting recognition for the past few years.  ...  Different feature extraction and classification techniques used in Malayalam off-line handwriting recognition are pointed out. Then databases chosen for these purposes are also discussed.  ...  Guru 2007 1D Wavelet transform Multi Layer Perceptron (MLP) 4950 samples of 33 classes. 73.8% Lajish.V.L 2008 fuzzy-zoned normalized vector distance Class modular neural network.  ... 
doi:10.26637/mjm0901/0106 fatcat:af72tejz3rcofpdxky4x56jyrq

In-Situ Learning in Multi-net Systems [chapter]

Matthew Casey, Khurshid Ahmad
2004 Lecture Notes in Computer Science  
We examine collaboration in multi-net systems through in-situ learning, exploring how generalisation can be improved through the simultaneous learning in networks and their combination.  ...  Results for these are compared with existing approaches, demonstrating that in-situ trained systems perform better than similar pre-trained systems.  ...  For MONK 2, 3 and WBCD, the SLM system out-performs the other single network and multi-net systems.  ... 
doi:10.1007/978-3-540-28651-6_112 fatcat:dm5nqpqfvzgjjfeydc6kcipxa4

Context-Aware Fuzzy ArtMap for Received Signal Strength Based Location Systems

Uzair Ahmed, Andrey Gavrilov, Sungyoung Lee, Young-Koo Lee
2007 Neural Networks (IJCNN), International Joint Conference on  
Intrinsically RSS based positioning is a multi-class pattern recognition problem.  ...  We present a Context-aware Fuzzy ArtMap neural network that provides competitive location accuracy in comparison with modular approach while leveraging online and incremental learning capabilities to location  ...  to Fuzzy ArtMap based location system we implemented two other neural network models for our location system i) Multi Layer Perceptron (MLP) and ii) Learning Vector Quantization (LVQ), in order to evaluate  ... 
doi:10.1109/ijcnn.2007.4371392 dblp:conf/ijcnn/AhmadGLL07 fatcat:jhgxtww7ozcbrez5qzbaspiifq

Artificial neural networks for document analysis and recognition

S. Marinai, M. Gori, G. Soda
2005 IEEE Transactions on Pattern Analysis and Machine Intelligence  
Artificial neural networks have been extensively applied to document analysis and recognition.  ...  Index Terms-Character segmentation, document image analysis and recognition, layout analysis, neural networks, preprocessing, recursive neural networks, word recognition. ae . S. Marinai and G.  ...  They are also very grateful to the participants of the tutorial Artificial Neural Networks for Document Analysis and  ... 
doi:10.1109/tpami.2005.4 pmid:15628266 fatcat:oanhch2aufctfd2fly47337j2u

Invariant Object Recognition Using Circular Pairwise Convolutional Networks [chapter]

Choon Hui Teo, Yong Haur Tay
2006 Lecture Notes in Computer Science  
We compared the recognition accuracy and training time complexity between our approach and a benchmark generic object recognizer LeNet7 which is a monolithic convolutional network.  ...  Out of many techniques, convolutional network (CN) is proven to be a good candidate in this area.  ...  To combine these pairwise ratios into a final decision for the multi-class problem, a metaclassifier such as MLP can be trained to map them into their corresponding desired outputs.  ... 
doi:10.1007/978-3-540-36668-3_167 fatcat:666h5xydjravdhkhdd6xlvl73u
« Previous Showing results 1 — 15 out of 1,557 results