Dr. Edgar Roman-Rangel

Address

Viper group - CVM Lab.
Centre Universitaire d'Informatique - University of Geneva.
Researcher within the Viper group
E-mail: Edgar [dot] RomanRangel [AT] unige [dot] ch
Phone: +41 22 379 0037

Biography

I am a researcher in Computer Vision and Machine Learning, with interests in Cross-Modal Information Retrieval. I like working on applications in areas of Cultural Heritage and Digital Arts.

I obtained my PhD from the École Polytechnique Fédérale de Lausanne (EPFL) in 2013, while working as a research assistant at the Idiap Research Institute. Since 2013, I have been working with the Viper group at the University of Geneva, Switzerland. During 2014 and 2015 I visited the National Institute of Anthropology and History of Mexico, working in the Tepalcatl project funded by the Swiss-NSF. I have served as part of technical committee and program committee for several international journals and conferences.

Research Interests

  • Computer Vision.
  • Machine Learning.
  • Cross-Modal Information Retrieval.
  • Cultural Heritage.
  • Digital Arts.

Projects Involved

  • MAAYA @ Viper Group - UniGe
  • Tepalcatl @ National Institute of Anthropology and History of Mexico (INAH)
  • CODICES @ Idiap Research Institute
  • IM2 @ Viper Group - UniGe

Code and data

  • HoSO: Local shape descriptor for 3D surfaces.
  • Potsherds: Set of 207 3D surfaces scanned from external views of potsherds from the Teotihuacan culture.
  • Hoosc: Local shape descriptor for binary images (2D shapes).

Publications

PhD Thesis:

  • Edgar Roman-Rangel.
    "Statistical Shape Descriptors for Ancient Maya Hieroglyphs Analysis".
    École Polytechnique Fédérale de Lausanne (EPFL). 2012. [PDF].

Journals:

  • Edgar Roman-Rangel, Diego Jimenez-Badillo, and Stephane Marchand-Maillet.
    "Classification and Retrieval of Archaeological Potsherds using Histograms of Spherical Orientations".
    ACM Journal on Computing and Cultural Heritage (JOCCH). 9(3):Online first. September, 2016.
  • Edgar Roman-Rangel, Changhu Wang, and Stephane Marchand-Maillet.
    "SimMap: Similarity Maps for Scale Invariant Local Shape Descriptors".
    Neurocomputing. 175(B):888–898. January, 2016.
  • Jun Guo, Changhu Wang, Edgar Roman-Rangel, Hongyang Chao, and Yong Rui.
    "Building Hierarchical Representations for Oracle Character and Sketch Recognition".
    IEEE Transactions on Image Processing. 25(1):104–118. January, 2016.
  • Edgar Roman-Rangel and Stephane Marchand-Maillet.
    "Shape-based Detection of Maya Hieroglyphs using Weighted Bag Representations".
    Pattern Recognition. 48(4):1157–1169. April, 2015.
  • Edgar Roman-Rangel, Jean-Marc Odobez, and Daniel Gatica-Perez.
    "Assessing Sparse Coding Methods for Contextual Shape Indexing of Maya Hieroglyphs".
    Journal of Multimedia (JMM). 7(2):179–192. Special Issue in Recent Achievements in Multimedia for Cultural Heritage. April, 2012.
  • Edgar Roman-Rangel, Carlos Pallan, Jean-Marc Odobez, and Daniel Gatica-Perez.
    "Analyzing Ancient Maya Glyph Collections with Contextual Shape Descriptors".
    International Journal of Computer Vision (IJCV). 94(1):101–117. Special Issue in Cultural Heritage and Art Preservation. August, 2011. Best Idiap Paper Award 2010. Best Poster ENS/INRIA CVML Summer School, Paris 2011.

Conferences:

  • Edgar Roman-Rangel and Stephane Marchand-Maillet.
    "Indexing Mayan Hieroglyphs with Neural Codes".
    Accepted to the 23rd International Conference on Pattern Recognition (ICPR). Cancun, Mexico. December, 2016.
  • Stephane Marchand-Maillet, Edgar Roman-Rangel, Hisham Mohamed, and Frank Nielsen.
    "Quantifying the Invariance and Robustness of Permutation-based Indexing Schemes".
    In Proceedings of the 9th International Conference on Similarity Search and Applications (SISAP). Tokyo, Japan. October, 2016.
  • Edgar Roman-Rangel, Gulcan Can, Stephane Marchand-Maillet, Rui Hu, Carlos Pallan Gayol, Guido Krem- pel, Jakub Spotak, Jean-Marc Odobez, and Daniel Gatica-Perez.
    "Transferring Neural Representations for Low-dimensional Indexing of Maya Hieroglyphic Art".
    In Proceedings of the 14th European Conference on Computer Vision (ECCV), 3rd Workshop on Computer Vision for Art Analysis. Amsterdam, The Netherlands. October, 2016.
  • Edgar Roman-Rangel, Diego Jimenez-Badillo, and Stephane Marchand-Maillet.
    "Rotation Invariant Local Shape Descriptors for Classification of Archaeological 3D Models".
    In Proceedings of the 8th Mexican Conference on Pattern Recognition (MCPR). Guanajuato, Mexico. June, 2016.
  • Diego Jimenez-Badillo and Edgar Roman-Rangel.
    "Application of the ‘Bag of Words’ model (bow) for analysing archaeological potsherds".
    CAA2015. Keep the Revolution Going. Proceedings of the 43rd. Annual Conference on Computer Applications and Quantitative Methods in Archaeology. Siena, Italy. March-April, 2016.
  • Diego Jimenez-Badillo, Salvador Ruiz-Correa, Edgar Roman-Rangel, Rogelio Hasimoto-Beltran, Mario G. Canul Ku, and Omar Mendoza-Montoya.
    "A computer platform to analyse the shape of museum objects in 3D".
    In Proceedings of the International Symposium on VIRTUAL ARCHAEOLOGY Museums & Cultural Tourism. Delphi, Greece. September, 2015.
  • Edgar Roman-Rangel and Diego Jimenez-Badillo.
    "Similarity Analysis of Archaeological Potsherds using 3D Surfaces".
    In Proceedings of the 7th Mexican Conference on Pattern Recognition (MCPR). Mexico City, Mexico. June, 2015.
  • Edgar Roman-Rangel and Stephane Marchand-Maillet.
    "Automatic Removal of Visual Stop-Words".
    In Proceedings of the ACM International Conference in Multimedia (ACM-MM). Orlando, USA. November, 2014.
  • Edgar Roman-Rangel, Diego Jimenez-Badillo, and Estibaliz Aguayo-Ortiz.
    "Categorization of Aztec Potsherds using 3D Local Descriptors".
    In Proceedings of the 12th Asian Conference on Computer Vision (ACCV), 3rd Workshop on e-Heritage. Singapore. November, 2014.
  • Daniel Gatica-Perez, Carlos Pallan Gayol, Stephane Marchand-Maillet, Jean-Marc Odobez, Edgar Roman-Rangel, Guido Krempel, and Nikolai Grube.
    "The MAAYA Project: Multimedia Analysis and Access for Documentation and Decipherment of Maya Epigraphy".
    Workshop on Digital Humanities 2014 (DH2014). Lausanne, Switzerland. July, 2014.
  • Edgar Roman-Rangel and Stephane Marchand-Maillet.
    "HOOSC128: A More Robust Local Shape Descriptor".
    In Proceedings of the 6th Mexican Conference on Pattern Recognition (MCPR). Cancun, Mexico. June, 2014.
  • Edgar Roman-Rangel and Stephane Marchand-Maillet.
    "Bag-of-Visual-Phrases via Local Contexts".
    In Proceedings of the 2nd Asian Conference on Pattern Recognition (ACPR), Workshop on Recent Advances in Computer Vision and Pattern Recognition (RACVPR). Okinawa, Japan. November, 2013.
  • Edgar Roman-Rangel and Stephane Marchand-Maillet.
    "Stopwords Detection in Bag-of-Visual-Words: The Case of Retrieving Maya Hieroglyphs".
    In Proceedings of the 17th International Conference on Image Analysis and Processing (ICIAP), 2nd International Workshop on Multimedia for Cultural Heritage (MM4CH). Naples, Italy. September, 2013.
  • Edgar Roman-Rangel, Jean-Marc Odobez, and Daniel Gatica-Perez.
    "Evaluating Shape Descriptors for Detection of Maya Hieroglyphs".
    In Proceedings of the 5th Mexican Conference on Pattern Recognition (MCPR). Queretaro, Mexico. June, 2013.
  • Edgar Roman-Rangel, Carlos Pallan, Jean-Marc Odobez, and Daniel Gatica-Perez.
    "Searching the Past: An Improved Shape Descriptor to Retrieve Maya Hieroglyphs".
    In Proceedings of the ACM International Conference in Multimedia (ACM-MM). Scottsdale, USA. November - December, 2011.
  • Daniel Gatica-Perez, Edgar Roman-Rangel, Jean-Marc Odobez, and Carlos Pallan.
    "New world, New Worlds: Visual Analysis of Pre-Columbian Pictorial Collections".
    In Proceedings of the 1st International Workshop on Multimedia for Cultural Heritage (MM4CH). Modena, Italy. April, 2011.
  • Edgar Roman-Rangel, Carlos Pallan, Jean-Marc Odobez, and Daniel Gatica-Perez.
    "Retrieving Ancient Maya Glyphs with Shape Context".
    In Proceedings of the 12th IEEE International Conference on Computer Vision (ICCV), Workshop on eHeritage and Digital Art Preservation. Kyoto, Japan. September - October, 2009.
  • Juan Frausto-Solis, E. F. Roman, David Romero, Xavier Soberon, and Ernesto Liñán-García.
    "Analytically Tuned Simulated Annealing Applied to the Protein Folding Problem".
    In Proceedings of the International Conference on Computational Science. Beijing, China. May, 2007.