Abstract: Character Recognition is one of the important tasks in Pattern Recognition. The complexity of the character recognition problem depends on the character set to be recognized. In this paper it is developed 0ff-line strategies for the isolated handwrittenMarathi numerical (0 to ९ ) with Neurofuzzy logic has been provided. The neural fuzzysystem is considered for soft computing. This method improves the character recognition method. NeuroFuzzySystem is integration of Neural Network and Fuzzy logic. In that we are using neural fuzzysystem for classification.
In the survey of literature, we found that many researchers has done work to words the handwritten Devanagari characters (HDC). The research work on character recognition of Devanagari script was started in 1970, where Sinha R. K. and Mahabala  presented a syntactic pattern analysis system for the recognition of Devanagari character (DC). The first research on handwritten Devanagari character was published in Sethi I. K. and Chatterjee B. . After that researchers started working on the recognition of handwritten Devanagari characters and tried to solve the problem associated with them. Zoning based approach for printed kannada numeral recognition was proposed by Ravinda Hegadi , in which the features such as the zones where the end points stretch out and the regions that each numeral generates were used to recognize the Kannada numerals.
Many tedious tasks can be made more efficient by automating the process of reading handwrittennumerals. In such system an optical scanner converts each handwritten numeral to a digital image, and computer software classifies the image as one of the digits zero through nine. By reducing the need for human interaction, numeral-recognition systems can speed up jobs such as reading income tax returns, sorting inventory, and routing mail. Several steps are necessary to achieve this. A recognitionsystem must first capture digital image of handwrittennumerals. Before attempting to classify the numerals, some preprocessing image might be necessary. An algorithm must then classify each handwritten numeral as one of the ten decimal digits [8, 9].
The recognition of numeral digits is split into two main streams; offline numeral recognition and automatic real-time numeral recognition; commonly referred to as Online Numeral Recognition. Offline Arabic and Persian numeral recognition has received a lot of attention from the research community while online numeral recognition is still in its early stage of development and only a little research addresses this concept. This is due to many reasons some of which could be that the lack of comprehensive benchmarks for Arabic and Persian digits, writing on a digital platform which is not as accurate as writing on paper and the larger set of possible handwriting samples per digit. In offline recognition, the image of the complete sample is available from the outset while in online recognition the two- dimensional coordinates of the consecutive points of writing are stored based on their order. This implies that offline recognition uses spatio-luminance of an image for analysis and recognition, whereas, online recognition uses spatio- temporal representation of the input for analysis and recognition . The authors in  proposed a method for online handwritten Arabic numeral recognitionbased on fuzzy modeling. This method automatically generated the fuzzy models of the Arabic digits using the segments’ directions of Arabic online digits from the training set. Furthermore, they generate weights for the different segments automatically using the training sets. They used a two-phase approach for classification thus leveraging the Support Vector Machine (SVM) in the first phase and a fuzzybased
The aim of handwritten numeral recognition (HNR) system is to classify input numeral as one of K classes. Conventional HNR systems have two components: feature analysis and pattern classification, as shown in Fig. 1. In Feature analysis step, information relevant for pattern classifier. The pattern classification step labels the numeral as one of K classes using the class models. Over the years, considerable amount of work has been carried out in the area of HNR. Various methods have been proposed in the literature for classification of handwrittennumerals. These include Hough transformations, histogram methods, principal component analysis, and support vector machines, nearest neighbor techniques, neural computing and fuzzybased approaches. An extensive survey of recognition performance for large handwritten database through many kinds of features and classifiers is reported. In comparison with HNR systems of various non Indian scripts [e.g. Roman, Arabic, and Chinese] we find that the recognition of handwrittennumerals for Indian scripts is still a challenging task and there is spurt for work to be done in this area. A brief review of work done in recognition of handwrittennumerals written in Devanagari script is given below.
After acquiring the output from the develop diagnostic system for renal or kidney cancer by using adaptive neuro- fuzzy inference system(ANFIS), now the specialist doctors or expert detects the obtained output by comparing it with the actual or target output. Here, the target output is that output which is evaluated by the expertise for a particular inputs and same inputs are given to diagnostic system. After the comparing between obtained and target results, it is noticed that the developed diagnostic system for renal cancer gives the similar output as the target outputs. The performance of the diagnostic system for the renal cancer is calculated by considering the various parameters. These parameters are sensitivity, specificity, precision and classification accuracy. The table 1 shows the performance of developed diagnostic system for renal cancer by using adaptive neuro-fuzzy inference system by considering the various parameters.
accommodate them into a lexicondrivenrecognition approach where a large vocabulary is employed? In this paper, we focus our attention on the problem of matching a sequence of observations generated from high-level features extracted from words andstatistical models of characters (HMMs)in an efficient manner.They presented a paper in which a grading system for Punjabi writers based on offline handwritten Gurumukhi characters recognition. They proposed four feature extraction methods, namely, zoning, diagonal, directional, intersection and open end points and Zernike moments feature. For classification, k-NN, HMM and Bayesian classifiers are used. They also compare the handwriting of one writer with other writers. This approach can also be extended for other Indian scripts such as Bengali, Tamil and Devanagri 
Abstract— Arabic handwritten digit recognition is the science of recognition and classification of handwritten Arabic digits. It has been a subject of research for many years with rich literature available on the subject. Handwritten digits written by different people are not of the same size, thickness, style, position or orientation. Hence, many different challenges have to overcome for resolving the problem of handwritten digit recognition. The variation in the digits is due to the writing styles of different people which can differ significantly. Automatic handwritten digit recognition has wide application such as automatic processing of bank cheques, postal addresses, and tax forms. A typical handwritten digit recognition application consists of three main stages namely features extraction, features selection, and classification. One of the most important problems is feature extraction. In this paper, a novel feature extraction approach for off-line handwritten digit recognition is presented. Wavelets-based analysis of image data is carried out for feature extraction, and then classification is performed using various classifiers. To further reduce the size of training data-set, high entropy subbands are selected. To increase the recognition rate, individual subbands providing high classification accuracies are selected from the over-complete tree. The features extracted are also normalized to standardize the range of independent variables before providing them to the classifier. Classification is carried out using k-NN and SVMs. The results show that the quality of extracted features is high as almost equivalently high classification accuracies are acquired for both classifiers, i.e. k-NNs and SVMs.
Abstract— Biometrics plays a very crucial role in various pattern recognitions. Recognition systems are used for offline application and also for online applications. In the biometric family, palmprint based identification system has become one of the active research topics. Palmprint identification system as two phases one is the feature extraction phase and other is the identification phase. The purpose of this paper is to use adaptive neurofuzzy inference system for the identification phase. The objective of ANFIS is to integrate the best features of fuzzy systems and neural networks. An ANFIS based identification system is described here which uses palmprint as input. Experiments are carried out using the samples. Obtained results show that the system is reliable when considering it as a part of the verification mechanism.
In this paper, we proposed a new system to recog- nize unconstrained handwritten digit strings. We used a segmentation-recognition strategy for handwritten con- nected digits based on structural features and the Fuzzy- artificial immune system. First, we combined the back- ground and foreground analysis for extracting the feature points. For the background features, we applied a thinning procedure to the vertical projection profile of the image. For the foreground features, we applied a thinning proce- dure on the connected component and their edge. These feature points are linked to generate the possible segmen- tation paths in connecting digits. The resulted candidate segmentation paths are evaluated for removing the useless among them and keeping the best. The evaluation process is based on two main constraints. The first one is related to the features points of the candidate segmentation paths and the second one is related to its height. Finally, we in- troduced the Fuzzy-AIS classifier for ranking all possible segmentation paths and considering the best of them as the
Computer is a wonderful invention for the society. Since its invention, it has not only touched every aspect of human activities, but also has tremendously enhanced its own capability to fulfill the human requirements with greater ease and better accuracy. Automatic analysis of handwriting or hand written symbols/characters is one of the human requirements that have been a subject of intensive research for the last few decades and it is still far from the perfection. Such requirements gave birth to novel area of knowledge which is termed as machine learning. Machine learning describes the study of machines that can adapt to their environment and learn from observations. One such example is the conversion of handwritten documents into computer manipulable form and it is such an area which is still unexplored and need researcher‟s
CONCLUSION: An attempt is made to apply new technique for feature extraction. The method is promising one. The recognition rate is increased by reducing the features by nearly 23% and hence it is remarkable. Some researchers used these features for object detection. In this work, as per the information available, this is first time these features are used for character recognition. The obtained average result is on testing is 95.83% and on cross validation is 95.82%. MLP classifiers accuracy is 95.45% on testing and 95.32 on cross validation with Tanhaxon as Transfer function. The handwritten character recognition accuracy is better by using SVM classifier but the results of MLP classifiers are also comparable.
ABSTRACT: Cancer is a disease which causes increasing of human death rate in each year. Cancer can be occurred anywhere in human body, normally. Even after being involvement of several tests, it is very difficult to obtain the ultimate diagnosis, also for medical expertise due to of lot of different parameters obtained from the subject to be diagnosed. A major class of problems in medical sciences involves the diagnosis of disease based upon various tests performed upon patient. Over the past few decades, this has given rise to computerized diagnostic tools, intended to aid physicians in making sense out of the confusion of data. Clinical oncologists make diagnostic decisions about Cancer patient based on past professional experiences and knowledge, intelligent techniques are possibly the only class of automatic technique powerful enough to emulate the expert’s choice. Due to their stable behavior in the presence of noise in precision and uncertainty, the ANN and ANFIS techniques could potentially obtain better results than classical methods.
Hausa being the language spoken only among Hausa people, only limited research has been carried out in Hausa speech recognition in comparison to other Africa languages. Hausa has 27 letters (22 consonants and 5 vowels). Hausa language is one of the largest ethnic groups in Africa the Hausa are a diverse but culturally homogeneous people based primarily in the largest population of Hausa are concentrated in Nigeria and Niger. Voice recognition is the system by which sounds, words or phrases spoken by humans are converted into electrical signals and these signals are transformed into coding patterns to which meaning has been assigned. To extract valuable information from the voice and speech signal, make decisions on the process, and obtain results, the data needs to be manipulated and analyzed. This research work presents the achievability of MFCC to extract features and DTW to compare the Hausa words and numerals test patterns. The extraction and matching Process is implemented right after the Pre Processing or filtering signal is performed.
The autonomous mobile robot uses ultra-sonic sensors for detecting targets and avoiding collisions. The control system is organized in a top-bottom hierarchy of various tasks, commands, and behaviours. When multiple low-level behaviours are required, command fusion is used to combine the output of several neuro-fuzzy sub- systems. A switching coordination technique selects a suitable behaviour from the set of possible higher level behaviours. A parallel ( Transputers based ) fuzzy control is implemented for the robot guidance and obstacle avoidance. The mobile robot used in this work has been designed and constructed by the author at the University of Bahrain. The key issue of this research frame work is the utilization of a neurofuzzy system that runs over a parallel Trasputers. This has shown the ability to reduce the computational time needed for the movement.
The basic Sugeno_type fuzzy model is used which is a single network to filter the speckle noises . The input parameters are the fuzzy values based on the difference between the main pixel and its neighboring window pixels. The window size is number of input layers nodes in the system. This system uses a 5×5 window sizes hence five input nodes. Each node of the input layer is coupled with its neighboring window pixel and therefore the data used in this layer are fuzzy data. The hidden layers in the system provide knowledge to the systembased on the fuzzy rules and their implications . The inferences to the system are based on the fuzzy IF-THEN rules, which involve the parameters of the system. The weights are added to the network between the input layers and the hidden layers which are binary values. To improve the efficiency of the encoding system a set of five binary weights which identify the pattern of pixels is used. With this encoding mechanism a five bit substring is evolved. These 5-bit substrings result in three patterns of 90, 180 and 270 degree rotations. Optimization is done to the non-zero elements in weight-sets which identify a pattern in the neighboring window . Binary weights in the genetic string are optimized in training steps. Estimation of the noise amplitude in the neighboring patterns is applied the same manner as applying the local statistics .
In current years, recognition of handwriting has been a subject of interest in the area of image classification and pattern recognition. Character recognition can be employed in typewritten, printed, or handwritten characters. Due to a huge diversity in the writing patterns of people, recognition and classifications of characters is more complex.
Projek penyelidikan ini memperkenalkan satu pendekatan yang sistematik untuk reka bentuk satu Fuzzy Inference System (FIS) berdasarkan satu kelas Rangkaian Neural (Neural Network) bagi menilai prestasi pelajar-pelajar. Sistem kabur telah mencapai kejayaan dalam pelbagai aplikasi untuk menyelesaikan pelbagai masalah. Kini, gabungan sistem kabur dan rangkaian neural merupakan satu aplikasi yang berjaya dalam teknik komputeran lembut, ia mempunyai ciri-ciri hibrid dan keupayaan untuk belajar. Teknik yang menggunakan sistem fuzzy ini gabung dengan rangkaian neural untuk meningkatkan beberapa ciri-cirinya, seperti kelonggaran, kelajuan, dan kebolehsuaian, dinamakan adaptive neuro-fuzzy inference system (ANFIS). Pentaksiran dan penaakulan terhadap prestasi pelajar bukan satu tugas yang mudah, terutamanya apabila ia melibatkan banyak sifat ataupun faktor yang lain. Lagipun, pengetahuan diperolehi daripada pakar domain untuk menentukan kriteria pembelajaran pelajar dan juga keputusan tentang tahap penguasaan mereka. Memandangkan perolehan maklumat daripada pakar domain adalah tidak lengkap dan tidak pasti. Untuk mengatasi masalah ini, projek-projek ini akan menjalankan taakulan prestasi pelajar berdasarkan ANFIS. Kaedah ini boleh menghasilkan hasil-hasil berangka rangup untuk meramalkan prestasi pelajar itu. Keputusan-keputusan bagi pendekatan ANFIS akan dibandingkan dengan pendekatan FIS pakar manusia.
Adaptive neuro-fuzzy method (or Adaptive neuro-fuzzy inference system, ANFIS) has been became a popular method in control area. In this section, we give a brief description of the principles of Adaptive neuro-fuzzy inference system (ANFIS). The basic structure of the type of fuzzy inference system could be seen as a model that maps input characteristics to input membership functions. Then it maps input membership function to rules and rules to a set of output characteristics. Finally it maps output characteristics to output membership functions, and the output membership function to a single valued output or a decision associated with the output. It has been considered only fixed membership functions that were chosen arbitrarily. Fuzzy inference is only applied to only modeling systems whose rule structure is essentially predetermined by the user's interpretation of the characteristics of the variables in the model. However, in some modeling situations, it cannot be distinguish what the membership functions should look like simply from looking at data. Rather than choosing the parameters associated with a given membership function arbitrarily, these parameters could be chosen so as to tailor the membership functions to the input/output data in order to account for these types of variations in the data values. In such case the necessity of the adaptive neurofuzzy inference system becomes obvious. The parameters associated with the membership functions changes through the learning process. The computation of these parameters (or their adjustment) is facilitated by a gradient vector. This gradient vector provides a measure of how well the fuzzy inference system is modeling the input/output data for a given set of parameters. When the gradient vector is obtained, any of several