Search

Optical Character Recognition (OCR) is the process which enables a system to without human intervention identifies the scripts or alphabets written into the users’ verbal communication. Optical Character identification has grown to be individual of the mainly flourishing applications of knowledge in the field of pattern detection and artificial intelligence. In our survey we study on the various OCR techniques. In this paper we resolve and examine the hypothetical and numerical models of Optical Character Identification. The Optical character identification or classification (OCR) and Magnetic Character Recognition (MCR) techniques are generally utilized for the recognition of patterns or alphabets. In general the alphabets are in the variety of pixel pictures and it could be either handwritten or stamped, of any series, shape or direction etc. Alternatively in MCR the alphabets are stamped with magnetic ink and the studying machine categorize the alphabet on the basis of the exclusive magnetic field that is shaped by every alphabet. Both MCR and OCR discover utilization in banking and different trade appliances. Earlier exploration going on Optical Character detection or recognition has shown that the In Handwritten text there is no limitation lying on the script technique. Hand written correspondence is complicated to be familiar through due to diverse human handwriting style, disparity in angle, size and shape of calligraphy. An assortment of approaches of Optical Character Identification is discussed here all along through their achievement.

In this report, we present a method to extract the features and to evaluate the Japanese composition of Chinese Japanese leamers by the n-gram model. In our experiments, we propose to extract the features, and to try evaluation with four similarity indises by the 3-gram model among the 55 compositions written by Chinese students, EDR Japanese corpus for references of standard Japanese sentences, 37 compositions written by Japanese students, and the original of Japanese writers.

The objective of this work was the development of a temporal lobe epilepsy seizures characterization methodology. The characterization was realized through scalp EEG analysis.Recent researches, mainly pursued on signal recorded on depth electrodes, showed an evolution of the synchronizations between cerebral structures, allowing a characterization of the temporal lobe seizures dynamic.The originality of our work consisted in the extension of the methods developed on depth electrodes to the study of scalp EEGs. From the medical point of view, this work lied within the scope of the presurgical diagnosis assistance.Methods of relation measurement, such as coherence, Directed Transfer Function (DTF), linear (r²) or non-linear correlation (h²), were adapted to answer these problems. Various criteria, defined from clinical indications, allowed the description of the advantages of the non-linear correlation coefficient in the study of epileptic seizures from scalp EEGs.The characterization of the evolution of the non-linear correlation coefficient was used as the base of the development of three signal processing applications:- The first was the determination of the side Epileptic Zone (EZ) at the onset of a seizure.- The search for an epileptic pattern constituted the second application. The pattern was extracted by an algorithm that was looking for the similarities between two seizures.- A classification of the temporal lobe seizures constituted the third application. It was carried out by extracting a set of characteristics from the patterns extracted by the algorithm of second stage.The large data base, which contained eighty seven seizures recorded on forty-three patients (two crises per patient, three for one of them), guaranteed a statistical significativity.Concerning the results, a rate of good lateralization of about 88% was obtained. This result was very interesting, as in the literature it was some times reached, but only by exploiting a multimodal data set or with non-automatics methods. The classification gave a sensitivity of 85% for the medial seizures and 58% for the mésio-lateral ones., L'objectif de la thèse est le développement d'une méthode de caractérisation des crises d'épilepsie du lobe temporal à partir des EEG de surface et plus particulièrement de la zone épileptogène (ZE) à l'origine des crises.Des travaux récents validés en stéréoélectroencéphalographie (SEEG) ont démontré une évolution des synchronisations entre structures cérébrales permettant une caractérisation de la dynamique des crises du lobe temporal.L'originalité des travaux consiste à étendre les méthodes développées en SEEG, à l'étude des signaux EEG de surface. Du point de vue médical, ce travail s'inscrit dans le cadre de l'aide au diagnostic préchirugical.Des méthodes de mesure de relation, telles que la cohérence, la Directed Transfer Function (DTF), la corrélation linéaire (r²) ou la corrélation non-linéaire (h²), ont été adaptées pour répondre à cette problématique. Différents critères, définis à partir d'indications cliniques, ont permis la mise en évidence des avantages du coefficient de corrélation non-linéaire dans l'étude de l'épilepsie par les EEG de surface.L'exploitation de l'évolution du coefficient de corrélation non-linéaire est à la base de trois applications de traitement automatique du signal EEG :– La première est la détermination de la latéralisation de la ZE au départ d'une crise. Cette information constitue l'étape préliminaire lors de la recherche de la localisation de la ZE.– La recherche d'une signature épileptique constitue la seconde application. La signature est extraite par un algorithme de mise en correspondance et de mesure de similarités en intra-patients.– Une classification des crises du lobe temporal constitue la troisième application. Elle est réalisée en extrayant un ensemble de caractéristiques des signatures trouvées par l'algorithme de l'étape 2.La base de données qui contient quarante-trois patients et quatre-vingt-sept crises (deux crises par patient, trois pour l'un d'entre eux) garantit une certaine significativité statistique.En ce qui concerne les résultats, un taux de bonne latéralisation de l'ordre de 88% est obtenu. Ce taux est très intéressant, car dans la littérature, il peut être quelques fois atteint, mais en exploitant des données multimodalités et avec des méthodes non-automatiques. A l'issue de la classification, 85% des crises mésiales ont été correctement classifiées ainsi que 58% des crises mésio-latérales.