This paper presents a method for recognizing hand configurations of the Brazilian sign language (LIBRAS) using 3D meshes and 2D projections of the hand. Five actors performing 61 different hand configurations of the LIBRAS language were recorded twice, and the videos were manually segmented to extract one frame with a frontal and one with a lateral view of the hand. For each frame pair, a 3D mesh of the hand was constructed using the Shape from Silhouette method, and the rotation, translation and scale invariant Spherical Harmonics method was used to extract features for classification. A Support Vector Machine (SVM) achieved a correct classification of Rank1 = 86.06% and Rank3 = 96.83% on a database composed of 610 meshes. SVM classification was also performed on a database composed of 610 image pairs using 2D horizontal and vertical projections as features, resulting in Rank1 = 88.69% and Rank3 = 98.36%. Results encourage the use of 3D meshes as opposed to videos or images, given that their direct, real time acquisition is becoming possible due to devices like LeapMotion R or high resolution depth cameras.

Abstract. Computer programming is a major topic for Computer Science students and professionals. This activity presents challenges with respect to monitoring the students' knowledge acquisition over the domain content. The learning progress is defined by principles and skills acquisition. This work proposes an approach for evaluating and monitoring students' skills development in the field of computer programming. The method is based on the mapping of the student's input into expertise features. The setting is organized around a belief model based on the concept of bayesian networks. We expect to provide the human tutor with a partially automated way for monitoring students' learning by means of an intelligent tutoring system coupled to a virtual learning environment.Resumo. O ensino de programação de computadores é um tópico de grande importância na formação de profissionais da Computação. Essa atividade apresenta desafios com relação ao acompanhamento do progresso dos estudantes frente ao conteúdo ministrado. Tal progresso pode ser definido pela aquisição de princípios seguida do desenvolvimento de perícias. Propõe-se nesta pesquisa uma abordagem para avaliação e acompanhamento da aquisição de perícias no domínio da programação de computadores. O método é baseado no mapeamento de entradas do aprendiz em características da perícia. A abordagem inclui conceitos de redes bayesianas. Espera-se, com isso, fornecer ao tutor humano uma via parcialmente automatizada para o monitoramento da aprendizagem por meio de um sistema tutor inteligente acoplado a um ambiente virtual de aprendizagem.

Abstract. During their university education, Computer Science students have to develop computer programming skills. This specific topic represents challenges for both the student and the teacher. One of the main challenges is related to the monitoring of student progress by the teacher. Currently there are artifices that contribute to facilitate this monitoring, to mention: intelligent tutors systems and student profile modeling. These methods, however, still require intensive, and even manual, workload in order to feed the student model. This paper proposes a mechanism for feeding an apprentice model based on the use of an Abstract Syntax Tree parser and presents a comparison of the results of the same with respect to a human evaluator, the parser showed to be promising in the tests performed.Resumo. O caminho a ser trilhado por alunos da área de Ciência da Computação possui como um de seus pilares a programação de computadores. Este tópico em específico remete a desafios tanto para o aluno quanto para o professor. Um dos principais desafios consiste no acompanhamento do progresso do aluno por parte do professor. Atualmente existem artifícios que contribuem para facilitar este acompanhamento, a citar: sistemas tutores inteligentes e métodos de modelagem do perfil do aprendiz. Estes métodos, por vez, ainda pecam no sentido de exigirem grande carga de trabalho para alimentação do modelo do aprendiz, muitas vezes feita de forma manual. Este trabalho propõe um mecanismo de alimentação para um modelo de aprendiz baseado no uso de um parser baseado em Árvores de Sintaxe Abstratas e apresenta um comparativo dos resultados do mesmo com relação a um avaliador humano, o parser se mostrou promissor nos testes realizados. IntroduçãoA programação de computadores é um dos fundamentos de todos os cursos da área de Ciência da Computação e, na maioria das vezes, apresenta sérios desafios tanto para o docente quanto para os alunos. Parte destes desafios se deve à dificuldade de mensurar o aprendizado e o progresso do aluno durante o processo. Por um lado o professor não possui formas precisas de avaliar os conhecimentos já adquiridos e, por outro lado, o

Contextualized in the teaching of computer programming in Computing courses, this research investigates aspects and strategies for automatic source code assessment. Continuous on-time assessment of source codes produced by students is a challenging task for teachers. The literature presents different methods for automatic assessment of source code, mostly focusing on technical aspects, such as functional correctness assessment and error detection. This paper presents the A-Learn EvId method, having as the main characteristic its focus on the assessment of high-level skills instead of technical aspects. Automatically assessing high-level skills gives insights into the thought process students used to elaborate their responses, contributing to quality and timely feedback generation. The method is characterized by three fundamental steps: (1) inserting students' source code as input data; (2) identifying evidence of skills through automatic strategies; and (3) representing identified skills through a student model. The following contributions are highlighted: updating the state of the art on the topic; a set of 37 skills identifiable through 9 automatic source code assessment strategies; construction of datasets totaling 8651 source codes.

During the computer programming learning process, the student's progress is marked by new skills acquisition. The monitoring of this progress, from the teacher's perspective, can be a complex task when performed from manual perceptions. The situation is even more complex when dealing with a large number of students simultaneously. One way of supporting this activity involves the use of student models, where abilities are mapped in order to facilitate the visualization of the concepts already acquired and the gaps to be completed. The use of this type of tool is beneficial, but it can still need a heavy workload when it requires the model manual feeding, especially because it is necessary to continuously update it. This paper proposes the use of automatic mechanisms, based in source codes analysis, followed by the evidences detection, as inputs for the student model. A set of experiments that demonstrate the viability of the method is presented. The test scenario is composed by a dynamic bayesian network student model and databases of source code in C language.

This research is contextualized in the teaching of computer programming. Continuous assessment of source codes produced by students on time is a challenging task for teachers. The literature presents different methods for automatic evaluation of source code, mostly focusing on technical aspects. This research presents the A-Learn EvId method, having as the main differential the evaluation of high-level skills instead of technical aspects. The following results are highlighted: updating the state of the art through systematic mapping; a set of 37 skills identifiable through 9 automatic source code evaluation strategies; construction of datasets totaling 8651 source codes.

Resumo: A falta de assistência tecnológica para a inclusão do surdo na sociedade contemporânea é uma realidade existente. Este trabalho consiste em realizar a classificação das configurações de mão da Língua Brasileira de Sinais (LIBRAS), mediante Rede Neural Artificial Kohonen, que se baseia em aprendizado competitivo, simulando processos específicos do cérebro humano. O sistema é composto pelas seguintes etapas: extração de características das configurações de mão (imagens), treinamento da Rede Neural Artificial, testes e análise dos resultados.