Project

Back to overview

Tepalcatl

English title Automatic Categorization of Potsherds
Applicant Roman Rangel Edgar
Number 152166
Funding scheme Early Postdoc.Mobility
Research institution Instituto Nacional de Antropologia e Histori (INAH)
Institution of higher education Institution abroad - IACH
Main discipline Electrical Engineering
Start/End 01.04.2014 - 30.09.2015
Show all

All Disciplines (2)

Discipline
Electrical Engineering
Archaeology

Keywords (5)

cultural heritage; machine learning; potsherds; computer vision; image categorization

Lay Summary (French)

Lead
Tepalcatl veut dire tesson en Nahuatl, la langue parlée par les Aztèques et des autres civilisations précolombiennes qui habitaient les régions du centre et du sud du Mexique actuel. Le projet Tepalcatl étudiera le potentiel d'application des méthodes informatiques pour aider à la catégorisation de tessons de poterie de valeur archéologique.
Lay summary

Objectifs

Il y a deux objectifs principaux: l'investigation de méthodes statistiques pour la reconnaissance visuelle de tessons de poterie; et le développement de méthodes informatiques pour aider les archéologues à classer ce matériel.

Contexte scientifique et sociétal

Les tessons de poterie sont le matériel le plus étudié dans les sites d'excavation, car ils fournissent des évidences pertinentes pour l'analyse historique et culturelle. Plus précisément, la classification de tessons de poterie permet de définir des typologies céramiques, c’est-à-dire des styles de céramiques qui servent comme bases pour l'identification de périodes et d'influences culturelles de civilisations anciennes. La catégorisation de tessons est donc un défi important qui demande des outils efficaces qui peuvent être utilisés in situ.

L'Institut National d'Anthropologie et d'Histoire de Mexique (INAH) a sous son aile une grande collection de plus de 17000 tessons de poterie de l'ancienne ville de Teotihuacan, qui était le premier centre urbain et la plus grande ville d'Amérique entre 100-600 apr. J.-C. Actuellement, l'INAH s'intéresse à l'application de méthodes informatiques pour aider la classification de ce matériel, ainsi que la future classification in situ de nouveau matériel.

En général, Tepalcatl est un effort bi-disciplinaire qui a l'ambition d'exploiter le potentiel des méthodes informatiques pour fournir des outils qui répondent aux besoins archéologiques.
Direct link to Lay Summary Last update: 27.01.2014

Responsible applicant and co-applicants

Publications

Publication
Application of the ‘Bag of Words’ model (bow) for analysing archaeological potsherds
Diego Jimenez-Badillo, Edgar Roman-Rangel (2016), Application of the ‘Bag of Words’ model (bow) for analysing archaeological potsherds, in CAA2015. Keep the Revolution Going. Proceedings of the 43rd. Annual Conference on Computer Applicati, Siena, Italy.Archaeopres, Oxford.
Rotation Invariant Local Shape Descriptors for Classification of Archaeological 3D Model
Edgar Roman-Rangel, Diego Jimenez-Badillo, Stephane Marchand-Maillet (2016), Rotation Invariant Local Shape Descriptors for Classification of Archaeological 3D Model, in Proceedings of the 8th Mexican Conference on Pattern Recognition (MCPR), Guanajuato, MexicoLecture Notes in Computer Science (LNCS) - Springer, Switzerland.
SimMap: Similarity Maps for Scale Invariant Local Shape Descriptors
Roman-Rangel Edgar, Wang Changhu, Marchand-Maillet Stephane (2016), SimMap: Similarity Maps for Scale Invariant Local Shape Descriptors, in Neurocomputing, 175(B), 888-898.
Similarity Analysis of Archaeological Potsherds using 3D Surfaces
Roman-Rangel Edgar, Jimenez-Badillo Diego (2015), Similarity Analysis of Archaeological Potsherds using 3D Surfaces, in 7th Mexican Conference on Pattern Recognition. Lecture Notes in Computer Science., Mexico City, Mexico.Lecture Notes in Computer Science (LNCS) - Springer., Switzerland..
Automatic Removal of Visual Stop-Words
Roman-Rangel Edgar, Marchand-Maillet Stephane (2014), Automatic Removal of Visual Stop-Words, in MM '14 Proceedings of the ACM International Conference on Multimedia, Orlando, USA.ACM, New York.
Categorization of Aztec Potsherds using 3D Local Descriptors
Roman-Rangel Edgar, Jimenez-Badillo Diego, Aguayo-Ortiz Estibaliz (2014), Categorization of Aztec Potsherds using 3D Local Descriptors, in Computer Vision - ACCV 2014 Workshops, Singapore Lecture Notes in Computer Science (LNCS) - Springer, Switzerland.
Classification and Retrieval of Archaeological Potsherds using Histograms of Spherical Orientations
Edgar Roman-Rangel, Diego Jimenez-Badillo, Stephane Marchand-Maillet, Classification and Retrieval of Archaeological Potsherds using Histograms of Spherical Orientations, in ACM Journal on Computing and Cultural Heritage (JOCCH).

Collaboration

Group / person Country
Types of collaboration
Stephane Marchand-Maillet, CVMLab - CUI. University of Geneva. Switzerland (Europe)
- in-depth/constructive exchanges on approaches, methods or results
- Publication
- Research Infrastructure
Rogelio Hasimoto Beltran, Computer Science - CIMAT. Mexico (North America)
- in-depth/constructive exchanges on approaches, methods or results

Scientific events

Abstract

The Tepalcatl project will focus on advancing the state-of-the-art in computational archaeology. More precisely, this project will investigate the potential of using clustering and classification methods, that provide stable quantitive results, to address the problem of automatic categorization of potsherds, which often requires qualitative analysis and expert knowledge. On the one hand, the Tepalcatl project has the technical objective of advancing the state-of-the-art in methods for efficient clustering and classification of image collections. On the other hand, it will provide tools to assist archaeologists in the automatic categorization of potsherds. The importance of this research project lays upon the fact that potsherds are the most studied material in any excavation site, as they provide with the relevant knowledge regarding the cultural and historical context for a given site, thus their categorization is an important challenge that requires efficient tools that can be used at excavation sites. Overall, this work represents a bi-disciplinary effort towards bringing the potential that computer science methods have in providing tools to satisfy archaeological needs.
-