Handwritten Digit Recognition: A Neural Network Demo

Digit recognition using single layer neural network with principal component analysis

Abstract: This paper presents an approach to digit recognition using single layer neural network classifier with Principal Component Analysis (PCA) The handwritten digit recognition is an important area of research as there are so many applications which are using handwritten recognition and it can also be applied to new application There are many algorithms applied to this computer vision problem and many more algorithms are continuously developed on this to make the handwritten recognition classify digits more accurately with less computation involved The proposed model in this paper aims to reduce the features to reduce computation requirements and successfully classify the digit into 10 categories (0 to 9) The system designed consists of backward propagation (BP) neural network and is trained and tested on the MNIST dataset of handwritten digit The proposed system was able to obtain 9839% accuracy on the MNIST 10,000 test dataset The Principal Component Analysis (PCA) is used for feature extraction to curtail the computational and training time and at the same time produce high accuracy It was clearly observed that the training time is reduced by up to 80% depending on the number of principal component selected We will consider not only the accuracy, but also the training time, recognition time and memory requirements for entire process Further, we identified the digits which were misclassified by the algorithm Finally, we generate our own test dataset and predict the labels using this system

Evolutionary Approaches to Rule Extraction from Neural Networks

Abstract: A short survey of existing methods of rule extraction from neural networks starts the chapter. Because searching rules is similar to NP-hard problem it justifies an application of evolutionary algorithm to the rule extraction. The survey contains a short description of evolutionary based methods, as well. It creates a background to show own experiences from satisfying applications of evolutionary algorithms to this process. Two methods of rule extraction namely: REX and GEX are presented in details. They represent a global approach to rule extraction, perceiving a neural network by the set of pairs: input pattern and response produced by the neural network. REX uses prepositional fuzzy rules and is composed of two methods REX Michigan and REX Pitt. GEX takes an advantage of classical crisp rules. All details of these methods are described in the chapter. Their efficiency was tested in experimental studies using different benchmark data sets from UCI repository. The comparison to other existing methods was made and is presented in the chapter.

Handwriting Digit Recognition With Fuzzy Logic

Abstract: To recognize handwriting digit is not a difficult task for human, but for a computer, it could be very difficult. This project implements the Fuzzy Logic system to recognize the handwriting digit. There are 3 constraints need to be considered here, they are: the real data were written with the same pen; the real data will be scanned into image data and then converted to BW mode with other software outside this project; program will read image data file instead of capturing with special device such as camera. Software is implemented in matlab. The design of fuzzy logic will use fuzzy logic editor. Before processing with Fuzzy algorithm, it needs to process the image then to get its features. Only simple image processing technique will be used. Feature extraction was made with a vertical and two horizontal lines. The position of crossing point between these lines with the image data will be a feature. These pre-processed data will be an input parameter for the fuzzy system. The fuzzy system has 7 inputs and 1 output with 57 rules. The average result of recognizing process is 80% after membership functions tuning. Abstract in Bahasa Indonesia : Bukan masalah yang rumit bagi manusia untuk mengenali angka yang ditulis oleh orang lain, tetapi tidak untuk computer. Paper ini akan membahas bagaimana menggunakan Fuzzy Logic untuk mengenali tulisan angka. Ruang lingkup paper ini dibatasi oleh 3 hal berikut: angka akan ditulis dengan menggunakan pena yang sama; hasil tulisan tersebut akan discan menjadi data mentah dan diubah menjadi gambar BW dengan menggunakan software di luar proyek ini, data ini akan langsung dibaca oleh program untuk dikenali (program tidak membaca data melalui kamera). Program dibuat dalam matlab dengan dibantu fuzzy logic editor. Untuk dapat mengenali data gambar tadi, harus dilakukan pengambilan informasi yang mewakili data tersebut (feature extraction). Cara yang dilakukan sangat sederhana, yaitu dengan menggunakan 2 garis horisontal dan 1 garis vertikal. Posisi titik potong antara garis-garis tersebut dengan angka merupakan data yang menjadi input untuk fuzzy logic. Fuzzy logic diimplementasikan dengan menggunakan 7 input, 1 output dan 57 aturan. Dari hasil percobaan, didapat bahwa kemampuan rata-rata program untuk mengenali angka adalah 80%. Hasil ini didapat setelah dilakukan tuning terhadap fuzzy logic. Kata kunci : fuzzy logic, pengenalan karakter, pengenalan tulisan tangan.

Handwritten signature recognition using departure of images from independence

Abstract: In this paper we propose a new Handwritten Signature Recognition Algorithm. The Algorithm is based on pixel-to-pixel relationship between Images. The Algorithms are based on extensive statistical analysis, Standard Deviation, variance and Theory of Cross-Correlation. This is an extension work of Handwritten Signature Identification. This Algorithm supports the application environment and we strongly believe that "User Recognition" could be a solid platform for future research and study based on statistics and probability theory.

Using domain-specific basic functions for the analysis of supervised artificial neural networks

Abstract: This thesis presents a generic neural network analysis method that utilizes domain-specific basic functions that are easy to interpret by the user and that can furthermore be used to optimize neural network systems. In general, the analysis consists in describing the internal functionality of the neural network in terms of domain-specific basic functions, functions that can be considered basic in the application domain of the neural network.

Self-Organizing Maps

Abstract: The Self-Organising Map (SOM) algorithm was introduced by the author in 1981. Its theory and many applications form one of the major approaches to the contemporary artificial neural networks field, and new technologies have already been based on it. The most important practical applications are in exploratory data analysis, pattern recognition, speech analysis, robotics, industrial and medical diagnostics, instrumentation, and control, and literally hundreds of other tasks. In this monograph the mathematical preliminaries, background, basic ideas, and implications are expounded in a manner which is accessible without prior expert knowledge.

Artificial neural networks: a tutorial

Abstract: Artificial neural nets (ANNs) are massively parallel systems with large numbers of interconnected simple processors. The article discusses the motivations behind the development of ANNs and describes the basic biological neuron and the artificial computational model. It outlines network architectures and learning processes, and presents some of the most commonly used ANN models. It concludes with character recognition, a successful ANN application.

Comparison of learning algorithms for handwritten digit recognition

Abstract: COMPARISON OF LEARNINGALGORITHMS FOR HANDWRITTEN DIGITRECOGNITIONY. LeCun, L. Jackel, L. Bottou, A. Brunot, C. Cortes,J. Denker, H. Drucker, I. Guyon, U. M uller,E. Sackinger, P. Simard, and V. VapnikBell Lab oratories, Holmdel, NJ 07733, USAEmail: yann@research.att.comAbstractThis pap er compares the p erformance of several classi er algorithmson a standard database of handwritten digits. We consider not only rawaccuracy, but also rejection, training time, recognition time, and memoryrequirements.1

Knowledge-based intelligent techniques in character recognition

Abstract: An Introduction to Character Recognition - L.C. Jain and B. Lazzerini Recognition of Handwritten Digits in the Real World by a Neocognitron - H. Shouno, K. Fukushima and M. Okada Recognition of Rotated Patterns Using Neocognitron - S. Satoh, J. Kunoiwa, H. Aso and S. Miyuke Soft Computing Approach to Hand-written Numeral Recognition - J. F. Baldwin, T. P. Martin, and O. Stylianidis Handwritten Character Recognition Using an MLP - F. Sorbello, G. A. M. Gioiello, and S. Vitabile Signature Verification Based on Fuzzy Genetic Algorithm - J. N. K. Liu, and G. S. K. Fung Application of a Generic Neural Network to Handwritten Digit Classification - D. S. Banarse and A. Duller High-speed Recognition of Handwritten Amounts On Italian Checks - B. Lazzerini, L. M. Reyneri , F. Gregoretti, and A. Mariani Off-line Handwritten Word Recognition Using Hidden Markov Models - A. El-Yacoubi, R. Sabourin, M. Gilloux and C. Y. Suen Off-line Handwriting Recognition with Context Dependent Fuzzy Rules - A. Malaviya, F. Ivancic, J. Balasubramaniam and L. Peters License-plate Recognition - M. H. Brugge, J. A. G. Nijihuis, L. Spaanenburg, and J. H. Stevens Index