Roberto Marcondes Cesar Junior

Detection and Tracking of Facial Features in Video Sequences

This work presents a real time system for detection and tracking of facial features in video sequences. Such system may be used in visual communication applications, such as teleconferencing, virtual reality, intelligent interfaces, human-machine interaction, surveillance, etc. We have used a statistical skin-color model to segment face-candidate regions in the image. The presence or absence of a face in each region is verified by means of an eye detector, based on an efficient template matching scheme . Once a face is detected, the pupils, nostrils and lip corners are located and these facial features are tracked in the image sequence, performing real time processing.

ScriptLattes: an open-source knowledge extraction system from the Lattes platform

The Lattes platform is the major scientific information system maintained by the National Council for Scientific and Technological Development (CNPq). This platform allows to manage the curricular information of researchers and institutions working in Brazil based on the so called Lattes Curriculum. However, the public information is individually available for each researcher, not providing the automatic creation of reports of several scientific productions for research groups. It is thus difficult to extract and to summarize useful knowledge for medium to large size groups of researchers. This paper describes the design, implementation and experiences with scriptLattes: an open-source system to create academic reports of groups based on curricula of the Lattes Database. The scriptLattes system is composed by the following modules: (a) data selection, (b) data preprocessing, (c) redundancy treatment, (d) collaboration graph generation among group members, (e) research map generation based on geographical information, and (f) automatic report creation of bibliographical, technical and artistic production, and academic supervisions. The system has been extensively tested for a large variety of research groups of Brazilian institutions, and the generated reports have shown an alternative to easily extract knowledge from data in the context of Lattes platform. The source code, usage instructions and examples are available at http://scriptlattes.sourceforge.net/.

Towards effective planar shape representation with multiscale digital curvature analysis based on signal processing techniques

Abstract This paper presents and discusses a new approach to multiscale curvature analysis of digital contours that is based on digital signal processing techniques. The shape contour is expressed in terms of two one-dimensional signals (linearized x - and y -coordinates) and the derivative theorem is applied as a means of obtaining an interesting expression relating contour curvature and the spectra of its parametrized x - and y -signals. Multiscale curvature analysis is achieved through Gaussian lowpass filtering and the “shrinkage” of the original signals implied by such a process is effectively circumvented by an energy-based compensation scheme, which has allowed accurate quantitative identification of the curvature value. The concept of curvegram is introduced and exemplified and the overall performance of the proposed technique for curvature estimation is formally assessed with respect to a contour defined in terms of B-splines. Application examples to synthetic and real images have been included.

Brazilian bibliometric coauthorship networks

The Brazilian Lattes Platform is an important academic/résumé data set that registers all academic activities of researchers associated with different major knowledge areas. The academic information collected in this data set is used to evaluate, analyze, and document the scientific production of research groups. Information about the interactions between Brazilian researchers in the form of coauthorships, however, has not been analyzed. In this article, we identified and characterized Brazilian academic coauthorship networks of researchers registered in the Lattes Platform using topological properties of graphs. For this purpose, we explored (a) strategies to develop a large Lattes curricula vitae data set, (b) an algorithm for identifying automatic coauthorships based on bibliographic information, and (c) topological metrics to investigate interactions among researchers. This study characterized coauthorship networks to gain an in‐depth understanding of the network structures and dynamics (social behavior) among researchers in all available major Brazilian knowledge areas. In this study, we evaluated information from a total of 1,131,912 researchers associated with the eight major Brazilian knowledge areas: agricultural sciences; biological sciences; exact and earth sciences; humanities; applied social sciences; health sciences; engineering; and linguistics, letters, and arts.

Feature Selection Based on Fuzzy Distances between Clusters: First Results on Simulated Data

Automatic feature selection methods are important in many situations where a large set of possible features are available from which a subset should be selected in order to compose suitable feature vectors. Several methods for automatic feature selection are based on two main points: a selection algorithm and a criterion function. Many criterion functions usually adopted depend on a distance between the clusters, being extremely important to the final result. Most distances between clusters are more suitable to convex sets, and do not produce good results for concave clusters, or for clusters presenting overlapping areas, in order to circumvent these problems, this paper presents a new approach using a criterion function based on a fuzzy distance. In our approach, each cluster is fuzzified and a fuzzy distance is applied to the fuzzy sets. Experimental results illustrating the advantages of the new approach are discussed.

Eigenfaces Versus Eigeneyes: First Steps Toward Performance Assessment of Representations for Face Recognition

The Principal Components Analysis (PCA) is one of the most successfull techniques that have been used to recognize faces in images. This technique consists of extracting the eigenvectors and eigenvalues of an image from a covariance matrix, which is constructed from an image database. These eigenvectors and eigenvalues are used for image classification, obtaining nice results as far as face recognition is concerned. However, the high computational cost is a major problem of this technique, mainly when real-time applications are involved. There are some evidences that the performance of a PCA-based system that uses only the region around the eyes as input is very close to a system that uses the whole face. In this case, it is possible to implement faster PCA-based face recognition systems, because only a small region of the image is considered. This paper reports some results that corroborate this thesis, which have been obtained within the context of an ongoing project for the development of a performance assessment framework for face recognition systems. The results of two PCA-based recognition experiments are reported: the first one considers a more complete face region (from the eyebrows to the chin), while the second is a sub-region of the first, containing only the eyes. The main contributions of the present paper are the description of the performance assessment framework (which is still under development), the results of the two experiments and a discussion of some possible reasons for them.

Towards Automatic Discovery of co-authorship Networks in the Brazilian Academic Areas

In Brazil, individual curricula vitae of academic researchers, that are mainly composed of professional information and scientific productions, are managed into a single software platform called Lattes. Currently, the information gathered from this platform is typically used to evaluate, analyze and document the scientific productions of Brazilian research groups. Despite the fact that the Lattes curricula has semi-structured information, the analysis procedure for medium and large groups becomes a time consuming and highly error-prone task. In this paper, we describe an extension of the script Lattés (an open-source knowledge extraction system from the Lattes platform), for analysing individuals Lattes curricula and automatically discover large-scale co-authorship networks for any academic area. Given some knowledge domain (academic area), the system automatically allows to identify researchers associated with the academic area, extract every list of scientific productions of the researchers, discretized by type and publication year, and for each paper, identify the co-authors registered in the Lattes Platform. The system also allows the generation of different types of networks which may be used to study the characteristics of academic areas at large scale. In particular, we explored the nodes degree and Author Rank measures for each identified researcher. Finally, we confirm through experiments that the system facilitates a simple way to generate different co-authorship networks. To the best of our knowledge, this is the first study to examine large-scale co-authorship networks for any Brazilian academic area.

Locating and Tracking Facial Landmarks Using Gabor Wavelet Networks

A new approach for locating and tracking facial landmarks in video sequences is introduced in this paper. Our method is based on Gabor wavelet networks, an effective technique that represents a discrete face template as a linear combination of 2D Gabor wavelet functions. This wavelet representation allows positioning of facial landmarks (e.g. eyes, nose and mouth), even in the presence of glasses, beard and different facial expressions. The feature tracking is robust to homogeneous illumination changes and affine deformations of the face image. Moreover, the tracking appraoch considers the overall geometry of the face, thus being robust to deformations such as eye blinking and smile, which is usually a critical situation to most local-based traditional methods.

Shape Analysis Using the Spectral Graph Wavelet Transform

The present work describes a framework for morphological characterization of galaxies based on the Spectral Graph Wavelet Transform. A galaxy image is sampled with a number of points randomly chosen, whose Delaunay triangulation results in an arbitrary graph. The average intensity value in a 5 × 5 vicinity of a pixel related to a graph vertex is assigned to the corresponding graph vertex. A weight inversely proportional to the photometric distance between each pair of vertices is assigned to the respective graph edge. The Spectral Graph Wavelet Transform is computed from this weighted graph with real-valued vertices yielding a high-dimensional feature vector, which is reduced to a two dimensional vector through Principal Component Analysis. The proposed framework has been assessed through two case studies, namely, the case study of analyzing (i) 2D binary images from shapes and preliminary results of (ii) 2D gray tone images from galaxies. The obtained results imply the suitability of this framework for the characterization of galaxies images.

Morphological Neuron Classification Based on Dendritic Tree Hierarchy

The shape of a neuron can reveal interesting properties about its function. Therefore, morphological neuron characterization can contribute to a better understanding of how the brain works. However, one of the great challenges of neuroanatomy is the definition of morphological properties that can be used for categorizing neurons. This paper proposes a new methodology for neuron morphological analysis by considering different hierarchies of the dendritic tree for characterizing and categorizing neuronal cells. The methodology consists in using different strategies for decomposing the dendritic tree along its hierarchies, allowing the identification of relevant parts (possibly related to specific neuronal functions) for classification tasks. A set of more than 5000 neurons corresponding to 10 classes were examined with supervised classification algorithms based on this strategy. It was found that classification accuracies similar to those obtained by using whole neurons can be achieved by considering only parts of the neurons. Branches close to the soma were found to be particularly relevant for classification.