Text size
  • Small
  • Medium
  • Large
  • Standard
  • Blue text on blue
  • High contrast (Yellow text on black)
  • Blue text on beige

    Classifying Ancient West Mexican Ceramic Figures Using Three-Dimensional Modelling and Machine Learning

    Electronic Visualisation and the Arts (EVA 2017)

    London, UK, 11 - 13 July 2017


    Julia Forrester-Sellers, Alfredo Velasco, Robert Pickering & J C Diaz



    This work involves the creation of a taxonomy of three-dimensional posture and gesture combinations represented by ceramic human figures associated with shaft and chamber tombs in the western coastal states of Nayarit, Jalisco, Michoacán and Colima of Mexico. Gordon Hewes (1966) proposed that posture and gesture of actual human subjects communicates a tremendous amount of information that was not being gathered on a consistent basis (Hewes 1966:106). While Hewes presented his paper on “The Domain Posture” fifty years ago, his observation rings true today. The three-dimensional taxonomy should demonstrate that posture and gesture are correlated with sexual characteristics and are not random, but rather, an illustration of a purposeful series of postures and gestures that transmit cultural meaning within the ancient constructs of West Mexican culture. There is an opportunity for anthropologists to begin seeing and evaluating and documenting the non-verbal communications made through posture and gesture both in artefacts crafted as figures, as in the case of the shaft tomb figures, and in modern cultural anthropology.

    The process of classification of the ceramic human figures according to posture and gesture takes three major steps: three-dimensional imaging, posture representation and finally classifying these representations. The three-dimensional imaging consists of creating three-dimensional models of the ceramic human figures. The posture representation consists of collecting measures of the fourteen major limb joints of the human body as they are represented in three-dimensional models of the ceramic figures. The last step is to use learning algorithms to classify the posture representation of the ceramic figures. The goal is to demonstrate the usefulness of the threedimensional imaging to accurately and consistently identify key posture/gesture combinations for comparison with other biological and cultural variables expressed by the figures.

    Initially, this project focuses on archaeological material located in the Gilcrease Museum collection in Tulsa, Oklahoma, but will also utilise other collections as points of comparison. While this work will identify patterns of posture and gesture within West Mexican ceramic shaft tomb figures, its application does not end there. With the utilisation of technology in the form of three-dimensional imaging, its application spans both archaeological material from other cultures and modern cultural anthropology. This technology can advance studies related to the cultural meaning of this important class of artefacts. From a musicological perspective, this technology could become a critical element in metadata collection to document collections.


    PDF Icon PDF Version of this Paper 691(kb)