RESUMEN
OBJECTIVES: An evaluation of the effectiveness of a new computational system proposed for automatic classification, developed based on a Siamese network combined with Convolutional Neural Networks (CNNs), is presented. It aims to identify endodontic technical errors using Cone Beam Computed Tomography (CBCT). The study also aims to compare the performance of the automatic classification system with that of dentists. METHODS: One thousand endodontically treated maxillary molars sagittal and coronal reconstructions were evaluated for the quality of the endodontic treatment and the presence of periapical hypodensities by three board-certified dentists and by an oral and maxillofacial radiologist. The proposed classification system was based on a Siamese network combined with EfficientNet B1 or EfficientNet B7 networks. Accuracy, sensivity, precision, specificity, and F1-score values were calculated for automated artificial systems and dentists. Chi-square tests were performed. RESULTS: The performances were obtained for EfficienteNet B1, EfficientNet B7 and dentists. Regarding accuracy, sensivity and specificity, the best results were obtained with EfficientNet B1. Concerning precision and F1-score, the best results were obtained with EfficientNet B7. The presence of periapical hypodensity lesions was associated with endodontic technical errors. In contrast, the absence of endodontic technical errors was associated with the absence of hypodensity. CONCLUSIONS: Quality evaluation of the endodontic treatment performed by dentists and by Siamese Network combined with EfficientNet B7 or EfficientNet B1 networks was comparable with a slight superiority for the Siamese Network. CLINICAL RELEVANCE: CNNs have the potential to be used as a support and standardization tool in assessing endodontic treatment quality in clinical practice.
Asunto(s)
Tratamiento del Conducto Radicular , Tomografía Computarizada de Haz Cónico Espiral , Humanos , Tomografía Computarizada de Haz Cónico/métodos , Atención Odontológica , Diente MolarRESUMEN
Vector Quantization (VQ) is a technique with a wide range of applications. For example, it can be used for image compression. The codebook design for VQ has great significance in the quality of the quantized signals and can benefit from the use of swarm intelligence. Initialization of the Linde-Buzo-Gray (LBG) algorithm, which is the most popular VQ codebook design algorithm, is a step that directly influences VQ performance, as the convergence speed and codebook quality depend on the initial codebook. A widely used initialization alternative is random initialization, in which the initial set of codevectors is drawn randomly from the training set. Other initialization methods can lead to a better quality of the designed codebooks. The present work evaluates the impacts of initialization strategies on swarm intelligence algorithms for codebook design in terms of the quality of the designed codebooks, assessed by the quality of the reconstructed images, and in terms of the convergence speed, evaluated by the number of iterations. Initialization strategies consist of a combination of codebooks obtained by initialization algorithms from the literature with codebooks composed of vectors randomly selected from the training set. The possibility of combining different initialization techniques provides new perspectives in the search for the quality of the VQ codebooks. Nine initialization strategies are presented, which are compared with random initialization. Initialization strategies are evaluated on the following algorithms for codebook design based on swarm clustering: modified firefly algorithm-Linde-Buzo-Gray (M-FA-LBG), modified particle swarm optimization-Linde-Buzo-Gray (M-PSO-LBG), modified fish school search-Linde-Buzo-Gray (M-FSS-LBG) and their accelerated versions (M-FA-LBGa, M-PSO-LBGa and M-FSS-LBGa) which are obtained by replacing the LBG with the accelerated LBG algorithm. The simulation results point out to the benefits of the proposed initialization strategies. The results show gains up to 4.43 dB in terms of PSNR for image Clock with M-PSO-LBG codebooks of size 512 and codebook design time savings up to 67.05% for image Clock, with M-FF-LBGa codebooks with size N=512, by using initialization strategies in substitution to Random initialization.
RESUMEN
Imaging examinations are of remarkable importance for diagnostic support in Dentistry. Imaging techniques allow analysis of dental and maxillofacial tissues (e.g., bone, dentine, and enamel) that are inaccessible through clinical examination, which aids in the diagnosis of diseases as well as treatment planning. The analysis of imaging exams is not trivial; so, it is usually performed by oral and maxillofacial radiologists. The increasing demand for imaging examinations motivates the development of an automatic classification system for diagnostic support, as proposed in this paper, in which we aim to classify teeth as healthy or with endodontic lesion. The classification system was developed based on a Siamese Network combined with the use of convolutional neural networks with transfer learning for VGG-16 and DenseNet-121 networks. For this purpose, a database with 1000 sagittal and coronal sections of cone-beam CT scans was used. The results in terms of accuracy, recall, precision, specificity, and F1-score show that the proposed system has a satisfactory classification performance. The innovative automatic classification system led to an accuracy of about 70%. The work is pioneer since, to the authors knowledge, no other previous work has used a Siamese Network for the purpose of classifying teeth as healthy or with endodontic lesion, based on cone-beam computed tomography images.
Asunto(s)
Tomografía Computarizada de Haz Cónico , Diente , Redes Neurales de la ComputaciónRESUMEN
The performance of signal processing systems based on vector quantization depends on codebook design. In the image compression scenario, the quality of the reconstructed images depends on the codebooks used. In this paper, alternatives are proposed for accelerating families of fuzzy K-means algorithms for codebook design. The acceleration is obtained by reducing the number of iterations of the algorithms and applying efficient nearest neighbor search techniques. Simulation results concerning image vector quantization have shown that the acceleration obtained so far does not decrease the quality of the reconstructed images. Codebook design time savings up to about 40% are obtained by the accelerated versions with respect to the original versions of the algorithms.
RESUMEN
In this study the spelling errors of the 'Dyslexic Sight Words - DSW' list are analyzed according to the semiological classification. The spelling errors were made by schoolchildren with and without dyslexia. The high number of inaccuracies observed in the writing of the Group with dyslexia (GD) was often related to the complexity of syllabic structures and orthographic irregularity. The syllabic structures, in addition to the consonant-vowel (CV) pattern, often pose challenges for all students as they move through the alphabetic writing phase, early in literacy. This classification provides an understanding of the characteristics of Natural Spelling and Arbitrary Spelling, providing support for the teaching-learning of words by dyslexic students and is also relevant for the design of Portuguese language teacher training policies. In the teaching of the orthographic norm, the success and error when writing words should be followed by a reflection (metaorthographic skill) and monitoring of learning, both on the part of the teacher and on the part of the learner, reinforcing the knowledge of spelling patterns that will be triggered as the student is exposed to the explicit formal teaching of spelling.
RESUMEN
Signal processing is a very useful field of study in the interpretation of signals in many everyday applications. In the case of applications with time-varying signals, one possibility is to consider them as graphs, so graph theory arises, which extends classical methods to the non-Euclidean domain. In addition, machine learning techniques have been widely used in pattern recognition activities in a wide variety of tasks, including health sciences. The objective of this work is to identify and analyze the papers in the literature that address the use of machine learning applied to graph signal processing in health sciences. A search was performed in four databases (Science Direct, IEEE Xplore, ACM, and MDPI), using search strings to identify papers that are in the scope of this review. Finally, 45 papers were included in the analysis, the first being published in 2015, which indicates an emerging area. Among the gaps found, we can mention the need for better clinical interpretability of the results obtained in the papers, that is not to restrict the results or conclusions simply to performance metrics. In addition, a possible research direction is the use of new transforms. It is also important to make new public datasets available that can be used to train the models.
RESUMEN
In this paper, the authors present the Cognitive LF-Ant protocol for emergency reporting in healthcare wireless sensor networks. The protocol is inspired by the natural behaviour of ants and a cognitive component provides the capabilities to dynamically allocate resources, in accordance with the emergency degree of each patient. The intra-cluster emergency reporting is inspired by the different capabilities of leg-manipulated ants. The inter-cluster reporting is aided by the cooperative modulation diversity with spectrum sensing, which can detect new emergency reporting requests and forward them. Simulations results show the decrease of average delay time as the probability of opportunistic access increases, which privileges the emergency reporting related to the patients with higher priority of resources' usage. Furthermore, the packet loss rate is decreased by the use of cooperative modulation diversity with spectrum sensing.
Asunto(s)
Lógica Difusa , Monitoreo Fisiológico/métodos , Telemedicina/métodos , Telemetría/métodos , Tecnología Inalámbrica , Análisis por Conglomerados , Redes de Comunicación de Computadores , Simulación por Computador , HumanosRESUMEN
Este estudo contempla a criação de uma lista de 60 palavras frequentemente escritas de modo incorreto por disléxicos. A proposta é inspirada na metodologia que utiliza listas de 'sight words' no ensino de palavras de alta frequência na língua inglesa. Participaram 60 escolares, 30 do Grupo Dislexia (GD) e 30 do Grupo sem Dislexia (GSD). Foi aplicado um questionário a 30 profissionais, que declararam 40 palavras frequentemente escritas de modo incorreto por disléxicos. Concebeu-se uma lista das 60 palavras mais citadas pelos profissionais. Todos os escolares realizaram um ditado dessas palavras. Foi constatada diferença no desempenho do GD comparado ao GSD ao escrever a lista de palavras. No GD as dificuldades persistem por mais tempo e há maior variedade de alterações ortográficas. A concepção da lista de palavras associada ao número alto de erros cometidos pelo GD pode oferecer subsídios para abordagens de intervenção no âmbito clínico e educacional.
This work addresses the creation of a list of 60 words often misspelled by dyslexics. The proposal is inspired by the methodology that uses 'sight words' lists in the teaching of high frequency words in the English language. Sixty students participated, 30 of the Dyslexia Group (GD) and 30 of the Without Dyslexia Group (GSD). A questionnaire was applied to 30 professionals, who declared 40 words frequently misspelled by dyslexics. A list of 60 words most cited by professionals was designed. All students performed a 60-word dictation. There was a difference in GD performance compared to GSD when writing the word list. In GD, the difficulties persist for a longer time and there is a greater variety of spelling changes. The design of the word list associated with the number of misspelled made by the GD may offer support for intervention approaches in clinical and educational scope.
RESUMEN
The particulate matter (PM) concentration has been one of the most relevant environmental concerns in recent decades due to its prejudicial effects on living beings and the earth's atmosphere. High PM concentration affects the human health in several ways leading to short and long term diseases. Thus, forecasting systems have been developed to support decisions of the organizations and governments to alert the population. Forecasting systems based on Artificial Neural Networks (ANNs) have been highlighted in the literature due to their performances. In general, three ANN-based approaches have been found for this task: ANN trained via learning algorithms, hybrid systems that combine search algorithms with ANNs, and hybrid systems that combine ANN with other forecasters. Independent of the approach, it is common to suppose that the residuals (error series), obtained from the difference between actual series and forecasting, have a white noise behavior. However, it is possible that this assumption is infringed due to: misspecification of the forecasting model, complexity of the time series or temporal patterns of the phenomenon not captured by the forecaster. This paper proposes an approach to improve the performance of PM forecasters from residuals modeling. The approach analyzes the remaining residuals recursively in search of temporal patterns. At each iteration, if there are temporal patterns in the residuals, the approach generates the forecasting of the residuals in order to improve the forecasting of the PM time series. The proposed approach can be used with either only one forecaster or by combining two or more forecasting models. In this study, the approach is used to improve the performance of a hybrid system (HS) composed by genetic algorithm (GA) and ANN from residuals modeling performed by two methods, namely, ANN and own hybrid system. Experiments were performed for PM2.5 and PM10 concentration series in Kallio and Vallila stations in Helsinki and evaluated from six metrics. Experimental results show that the proposed approach improves the accuracy of the forecasting method in terms of fitness function for all cases, when compared with the method without correction. The correction via HS obtained a superior performance, reaching the best results in terms of fitness function and in five out of six metrics. These results also were found when a sensitivity analysis was performed varying the proportions of the sets of training, validation and test. The proposed approach reached consistent results when compared with the forecasting method without correction, showing that it can be an interesting tool for correction of PM forecasters.
Asunto(s)
Monitoreo del Ambiente/métodos , Material Particulado/análisis , Contaminación del Aire/análisis , Simulación por Computador , Finlandia , Humanos , Modelos Teóricos , Redes Neurales de la Computación , Tamaño de la Partícula , Factores de TiempoRESUMEN
ABSTRACT Purpose: to present a digital version of the Rapid Automatized Naming (RAN) test, designed to be used as an auxiliary tool for early detection of reading problems in children. Methods: the application was designed to provide resources, such as: a) registration; b) information security with the online authentication system - individuals are only seen by the person who registered them and, c) history of the results. In order to know the impressions of speech therapists and professionals of children's education area about this digital version of RAN test, a public opinion survey was conducted with unidentified participants by Google Forms. Results: the application has an automatic speech recognition and allows registering and accessing the history of the application, including statistical evaluation of the results. The application features a responsive interface that adapts to the dimensions of the mobile device with soft colors, and offers a variety of text font options. The images of the objects of the digital version were elaborated by a designer. Conclusion: the application presents itself as a useful digital tool that assists professionals to evaluate, early, children with reading difficulties, highlighting the following elements: exclusive and appropriate images to children's universe and the function of storing the results of users tested.
RESUMO Objetivo: apresentar uma versão digital do teste de Nomeação Automática Rápida (RAN), construído para ser uma ferramenta auxiliar para detectar precocemente problemas de leitura em crianças. Métodos: o aplicativo foi elaborado visando fornecer recursos, como: a) cadastro de indivíduos; b) segurança da informação com o uso do sistema de autenticação online - cada indivíduo é visualizado apenas por quem o cadastrou e c) histórico de resultados. Para conhecer as impressões de fonoaudiólogos e profissionais da área da educação sobre esta versão digital do teste de RAN, foi realizada uma pesquisa de opinião pública com participantes não identificados por meio do Google Forms. Resultados: o aplicativo conta com reconhecimento automático de fala e permite registro e acesso ao histórico da aplicação, incluindo avaliação estatística de resultados. O aplicativo apresenta uma interface responsiva que se adapta às dimensões do aparelho móvel, com cores suaves, e oferece opções variadas de fontes de texto. As imagens dos objetos desta versão digital foram elaboradas por uma designer, sendo imagens comuns ao universo infantil. Conclusão: o aplicativo apresenta-se como uma ferramenta digital útil, que auxilia profissionais da área da educação infantil a avaliar precocemente crianças com dificuldades na leitura com destaque aos seguintes elementos: imagens exclusivas e adequadas ao universo infantil e a função de armazenamento de resultados dos usuários testados.
RESUMEN
ABSTRACT Purpose: to present a new application for mobile devices, referred to as Desembaralhando, for intervention in the problem of dyslexic children mirror writring. Methods: the development of the application is the result of a set of clinical and speech therapy information and experiences, which points out frequency of letter mirroring as a challenging problem in children with dyslexia. The application, developed in the light of the multisensory approach, was created by a multidisciplinary team of computer scientists, a game designer and a speech therapist, in order to meet users requirements, such as appropriate fonts and colors. Results: the activities stimulate phonological awareness skills from the association between images and words, audio aids, as well as an original function that is the rotational movement of letters b/d and a/e, which facilitates the perception of the visual layout of the letters. Conclusions: guidelines such as the choice of typography and interface colors appropriate to dyslexic children are used to favor intervention, in order to minimize the difficulties of these children regarding letters mirroring.
RESUMO Objetivo: apresentar um novo aplicativo desenvolvido para dispositivos móveis, denominado Desembaralhando, para a intervenção no problema do espelhamento de letras por crianças disléxicas. Métodos: o desenvolvimento do aplicativo é resultado de um conjunto de informações de experiências fonoaudiológicas, que apontam a ocorrência de espelhamentos de letras como um problema desafiador em crianças com dislexia. O aplicativo, desenvolvido à luz da abordagem multissensorial, foi criado por uma equipe multidisciplinar composta de cientistas da computação, designer de jogos e fonoaudióloga e foi projetado para atender requisitos sintonizados com o público-alvo, como fontes e cores apropriadas. Resultados: as atividades desenvolvidas estimulam as habilidades de consciência fonológica por meio da associação entre imagem e palavra, recursos de áudio, além de uma funcionalidade original, que é o movimento de rotação das letras b/d e a/e, que facilita a percepção do traçado visual das letras. Conclusões: diretrizes tais como a escolha da tipografia e cores de interface, adequadas a crianças disléxicas, são utilizadas para favorecer a intervenção para minimizar as dificuldades dessas crianças no âmbito do espelhamento de letras.
RESUMEN
ABSTRACT This paper aimed at reviewing the literature related to national and international research on spelling difficulties by dyslexics and identifying the intervention approaches performed with regard to this topic. An integrative review of the literature was carried out in order to answer the question: considering the domain of the orthography, one of the challenges frequently faced by dyslexics, how are studies on the relationship between dyslexia and spelling characterized? The research was carried out in PubMed platform, Scopus database and Portal de Periódicos CAPES/MEC. To search the articles, the following descriptors were used: "dislexia" or "dyslexia" with the free terms "ortografia" or "spelling". One aspect should be highlighted: some works indicate that difficulties in the spelling performance by dyslexics are not exclusively due to phonological processing failures - they are also secondary to alterations in orthographic processing. A challenge faced by dyslexics is to retain phonological information to use in writing new orthographic forms. Researchers suggest that intervention strategies include phonological, orthographic and lexical activities. It is observed that few studies have analyzed the difficulties that dyslexics face when dealing with new words, as well as writing, correctly, frequently used words in their own language.
RESUMO O objetivo deste estudo é revisar a literatura relacionada a pesquisas nacionais e internacionais sobre as dificuldades ortográficas enfrentadas por disléxicos e identificar as abordagens de intervenção na temática supracitada. Foi realizada uma revisão integrativa da literatura que buscou responder à pergunta: Considerando o domínio da ortografia um dos desafios frequentemente enfrentados por disléxicos, como se caracterizam os estudos a respeito da relação entre dislexia e ortografia? A pesquisa foi realizada na plataforma PubMed e nas bases de dados Scopus e Portal de Periódicos da Capes. Para pesquisar os artigos, foram utilizados da forma combinada os descritores "dislexia" ou "dyslexia" e os termos livres "ortografia" ou "spelling". Um aspecto deve ser destacado: alguns trabalhos apontam que as dificuldades no desempenho da escrita por disléxicos não são decorrentes, exclusivamente, de falhas no processamento fonológico - são também secundárias a alterações no processamento ortográfico. Um desafio enfrentado por disléxicos é reter as informações fonológicas para utilizar na escrita de novas formas ortográficas. Pesquisadores sugerem que as estratégias de intervenção contemplem atividades fonológicas, ortográficas e lexicais. Observa-se que poucos estudos analisaram as dificuldades que disléxicos apresentam para lidar com palavras novas, bem como escrever ortograficamente palavras frequentemente utilizadas em sua própria língua.
RESUMEN
RESUMO O objetivo deste estudo é apresentar uma revisão integrativa da literatura, contemplando artigos científicos publicados em periódicos nacionais e internacionais que abordam o uso das tecnologias da informação e da comunicação (TIC), tais como computador, tablets, iPads, mobile phones, e-readers, realidade virtual e ambiente virtual de aprendizagem, aplicadas à dislexia. A base de dados escolhida para este estudo foi constituída de artigos científicos publicados no período de 2010 a 2015, a partir das seguintes bases eletrônicas de dados: Science Direct/Elsevier, SciELO - Scientific Electronic Library Online, MedLine - Medical Literature Analysis and Retrieval e o Portal de Periódicos da CAPES. Foram selecionados para este estudo 21 artigos científicos, sendo 20 (95,23%) artigos internacionais e um (4,77%) artigo nacional. Os trabalhos contemplados, no presente estudo, em geral, visam a construção e aplicação de instrumentos tecnológicos que possam vir a minimizar as dificuldades do disléxico no âmbito da aprendizagem da leitura e da escrita. Em meio à escassez de artigos publicados no Brasil, verifica-se a necessidade de mais estudos sobre essa temática, tendo em vista os benefícios das TIC no âmbito da avaliação e intervenção em dislexia constatados em artigos internacionais.
ABSTRACT The current paper aims to present an integrative literature review of scientific articles published in national and international journals, approaching the use of information and communication technology (ICT) such as computers, tablets, iPad’s, mobile phones, e-readers, virtual reality and virtual learning environments, applied to dyslexia. The database consists of scientific articles published between 2010 and 2015, from the following platforms: Science Direct/Elsevier, SciELO - Scientific Electronic Library Online, MedLine - Medical Literature Analysis and Retrieval and Portal de Periódicos da CAPES. Scientific articles were selected, corresponding to 20 international (95.23%) papers and 1 national (4.77%) paper. The works considered in the present paper in general aim at the development and application of technological instruments that can minimize the difficulties by dyslexics in reading and writing learning. Due to shortage of articles published in Brazil, there is a need for more studies on this topic in view of the benefits of ICT within evaluation and intervention in dyslexia observed in international articles.
RESUMEN
O fato de os surdos utilizarem a lingua de sinais, que lhes promove naturalmente o acesso ao conhecimento em geral, leva a dificuldades para ler e escrever, tendo em vista que esta leitura e esta escrita são produzidas em uma lingua que não é a sua: a lingua padrão do país. Sendo as marcas de oralidade favorecedoras da coesão textual, a pergunta norteadora deste estudo foi: o fato de utilizar marcas de oralidade na escrita faz com que esta seja mais compreensível? Objetivou-se, portanto, investigar a compreensão, por ouvintes, do texto escrito por surdos, em função da ocorrência de tais marcas na escrita. Fragmentos de escrita com mais de uma marca, com uma marca e sem marca de oralidade, produzidos por surdos, além de fragmentos produzidos por ouvintes, foram apresentados a estudantes universitários ouvintes, que atribuíram escores aos fragmentos em função da compreensão do conteúdo. Foi verificado que os fragmentos referidos como de conteúdo mais compreendido foram os produzidos por ouvintes, seguidos dos com mais de uma marca de oralidade produzidos por surdos; os menos compreendidos foram os sem marca de oralidade produzidos por surdos. O estudo sugere que as marcas de oralidade promoveram a coesão textual dos surdos, facilitando a compreensão do conteúdo da sua escrita por ouvintes.
The use of sign language by the deaf, though a means of providing access to knowledge, offers some specific difficulties on reading/writing due to the impossibility on acquiring the written code of the official spoken language. Taking into account that some oral cues favor textual cohesion, the questionthis paper is mainly concerned with is whether the use of oral cues in writing favors comprehension as well. The aim of this research was to offer written texts produced by the deaf to the non deaf to see how the text was understood by these speakers. Some written fragments contained two or more oral cues, some with just one cue or with no cues produced by the deaf and some texts produced by the non deafwere offered to university hearing students who were asked to score the texts by means of levels of comprehension. The results showed that the answers favored the texts produced by the non deaf people followed by those with more than two oral cues produced by the deaf; the texts that offered difficulty for comprehension were those with no oral cues produced by the deaf. This paper suggests that the oral cuesbring cohesion to the texts produced by the deaf thus favoring the hearer text comprehension.