Arnaldo de Albuquerque Araujo


Bolsa de Produtividade CNPq: 1D

Áreas de Pesquisa:
Sala: 7331
Ramal: 5873
arnaldo@dcc.ufmg.br

Página pessoal     Lattes    Google scholar 


Informações resumidas do Currículo Lattes

Currículo Lattes atualizado em 17/08/2021

Nome em citações bibliográficas: ARAÚJO, A. de A.;Araújo, Arnaldo de A.;de Albuquerque Araújo, Arnaldo;Araújo, Arnaldo Albuquerque;De A. Araújo, A.;A. Araújo;ARAUJO, ARNALDO DE ALBUQUERQUE;A. ARAÚJO, ARNALDO;A. ARAUJO, ARNALDO DE;DE A. ARAUJO, A.;ARAUJO, ARNALDO DE A.;Arajujo, Arnaldo de A.;ARAÚJO, ARNALDO DE ALBUQUERQUE;ARAUJO, A.D.A.;DE A ARAUJO, A.;ARAUJO, A.A.;DE ALBUQUERQUE ARAUJO, A.;DE A., A. ARAUJO;ARAÚJO, A.A.;DE ALBUQUERQUE ARAUJO, ARNALDO;ARAUJO, A. DE A.;DE A. ARAÚJO, ARNALDO


Formação acadêmica

Doutorado em Engenharia Elétrica na Universidade Federal da Paraíba em 1987
Mestrado em Engenharia Elétrica na Universidade Federal da Paraíba em 1981
Graduado em Engenharia Elétrica C Grande na Universidade Federal da Paraíba em 1978

Projetos de pesquisa em andamento

2019 a AtualProjeto de Cooperação Internacional CAPES/COFECUB Edital 008/2018
Projeto de Cooperação Internaciona CAPES/COFECUB 933/19, Hierarchical Graph-based Analysis of Image, Video and Multimedia Data. Methods of image, video and multimedia analysis which use hierarchical representation aim to explore the visual representation as region-oriented scale-space. This representation is a set of graph representations at different detail levels in which the representation at finer levels are nested with respect to those at coarser levels producing a hierarchy of partitions. This kind of data structure has been successfully applied to remote sensing, object detection and human action recognition. Despite the several ways for computing hierarchies of partitions, developing efficient and effective methods is not an easy task due to the semantic information which is needed for a segmentation. In fact, the state-of-the-art segmentation methods depend on the good gradients to produce good results. Furthermore, non-hierarchical methods could produce good segmentation results thanks to the (dis)similarity measure used for merging two regions, which can be transformed, without loss of quality, into hierarchical methods which incorporate some new properties. Thus, the main goal of this study is to advance in the state-of-the-art on \textit{hierarchy of partitions} taking into account aspects of efficiency, quality, making hierarchical and interactivity, as well as the use of hierarchical information to help in the information extraction and the label propagation. Moreover, we will investigate hierarchical visualization of all, image, video and multimedia, by using countour saliency maps. Finally, we will explore the criteria for hierarchical comparison and for hierarchical combination taking into account their contour saliency maps and learning methods. The results of these studies will be used for solving several applications like human action recognition, pornography detection, image and video region labeling, multimedia label propagation, image and video inpainting, among others. We will build upon existing research and skills at ESIEE Paris, Université Gustave Eiffel, LIGM, IRISA Rennes, Grenoble INP, UNICAMP, PUC Minas and UFMG to develop collaborative work exploiting complementarily of these institutions. Validade: Jan 2019 a Dez 2022
Integrantes: Silvio Jamil Ferzoli Guimarães (coordenador), Arnaldo de Albuquerque Araújo, Laurent Najman, Jean Cousty, Hélio Pedrini, Alexandre Xavier Falcão, Guillaume Gravier, Yukiko Kenmochi, Jefersson Alex dos Santos, Zenilton Kleber Gonçalves do Patrocínio Júnior, Benjamin Perret, Simon Malinowski, Ewa Kijak, Mauro Dalla Mura, Jocelyn Chanussot.

Projetos de desenvolvimento em andamento

Veja todos os projetos no Currículo Lattes

Últimas publicações

Artigos em periódicos

Hierarchical segmentation from a non-increasing edge observation attribute
2020. PATTERN RECOGNITION LETTERS.
From video pornography to cancer cells: a tensor framework for spatiotemporal description
2020. MULTIMEDIA TOOLS AND APPLICATIONS.
Efficient algorithms for hierarchical graph-based segmentation relying on the Felzenszwalb-Huttenlocher dissimilarity
2019. INTERNATIONAL JOURNAL OF PATTERN RECOGNITION AND ARTIFICIAL INTELLIGENCE.
Combining pixel domain and compressed domain index for sketch based image retrieval
2017. MULTIMEDIA TOOLS AND APPLICATIONS.
A mid-level video representation based on binary descriptors: A case study for pornography detection
2016. Neurocomputing (Amsterdam).
Multi-Histogram Equalization Methods for Contrast Enhancement and Brightness Preserving
2007. IEEE Transactions on Consumer Electronics.
Video Segmentation Based on 2D Image Analysis
2003. Pattern Recognition Letters.
Sum of Absolute Grey Level Differences: An Edge-Preserving Smoothing Approach
1985. Electronics Letters.

Trabalhos completos em congressos

FASTensor: A tensor framework for spatiotemporal description
2019. Workshop of Theses and Dissertations - SIBGRAPI.
A study of observation scales based on Felzenswalb-Huttenlocher dissimilarity measure for hierarchical segmentation
2019. 21st International Conference on Discrete Geometry for Computer Imagery.
Hierarchy Based Salient Regions: a region detector based on hierarchies of partitions
2018. 23rd Iberoamerican Congress on Pattern Recognition.
Exploring deep-based approaches for semantic segmentation of mammographic images
2018. 23rd Iberoamerican Congress on Pattern Recognition.
Algorithms for hierarchical segmentation based on the Felzenszwalb-Huttenlocher dissimilarity
2018. International Conference on Pattern Recognition and Artificial Intelligence.
Combination techniques for hyperspectral image interpretation
2017. IGARSS 2017 2017 IEEE International Geoscience and Remote Sensing Symposium.
Digital color restoration from slide images which use the color target Kodak Q-13
2017. IS&T Archiving Conference.
Evaluation of combinations of watershed hierarchies
2017. 8th. International Symposium on Mathematical Morphology.
Development of an Open Source Software for Real Time Optical Motion Capture
2016. 2016 XVIII Symposium on Virtual and Augmented Reality (SVR).
Video similarity search by using compact representations
2016. the 31st Annual ACM Symposium.
Near-duplicate video detection based on an approximate similarity self-join strategy
2016. 2016 14th International Workshop on ContentBased Multimedia Indexing (CBMI).
A Boosting-Based Approach for Remote Sensing Multimodal Image Classification
2016. 2016 29th SIBGRAPI Conference on Graphics, Patterns and Images (SIBGRAPI).
Stochastic Hierarchical Watershed Cut Based on Disturbed Topographical Surface
2016. 2016 29th SIBGRAPI Conference on Graphics, Patterns and Images (SIBGRAPI).
Decreasing the Number of Features for Improving Human Action Classification
2016. 2016 29th SIBGRAPI Conference on Graphics, Patterns and Images (SIBGRAPI).
Star: A contextual description of superpixels for remote sensing image classification
2016. 21st Iberoamerican Congress on Pattern Recognition.
Suavização Espacial: Uma Revisão.
1990. Jornada EPUSP/IEEE em Computação Visual (Artigo Tutorial).

Resumos expandidos em congressos

Sistema computacional automático para análise de estruturas rúpteis, análise geomatemática e representações cartográficas
1997. Simpósio de Quantificação em Geociências.
Comparação de algoritmos de suavização espacial
1994. Simpósio Brasileiro de Computação Gráfica e Processamento de Imagens.
Image coding for storing historical documents in a database
1993. Picture Coding Symposium.
PHOTOPIX: A color image quantization system.
1993. Picture Coding Symposium.
Suavização espacial no ambiente KHOROS
1993. Simpósio Brasileiro de Computação Gráfica e Processamento de Imagens.

Resumos em congressos

Criação de um acervo digital da vida e obra do pintor Alberto da Veiga Guignard
2005. XIV Semana de Iniciação Científica da UFMG.
A hypermedia information system to manage information about restoration and conservation of cultural movable properties
2004. 2nd Latin American Web Congress and the 10th Brazilian Symposium on Multimedia and the Web.
Um Workflow de Digitalização para Geração do Acervo Digital do Cecor
2004. 2nd Latin American Web Congress and the 10th Brazilian Symposium on Multimedia and the Web.
Arcabouço para a Recuperação de Imagens Digitais de Fotografias Históricas Através da Utilização de Ontologias
2004. 1º Forum Brasileiro do Patrimônio Cultural.
Digitalização e Gerenciamento do Acervo sobre Conservação e Restauração de Bens Culturais Móveis do Cecor
2004. 1º Forum Brasileiro do Patrimônio Cultural.

Veja todas as publicações no Currículo Lattes

Orientações em andamento

Mestrado

Doutorado

Edward Jorge Yuri Cayllahua Cahuina. Algorithms for Hierarchical Segmentation. Início: 2014. Universidade Federal de Minas Gerais (Orientador principal)

Veja todas as orientações no Currículo Lattes