Automatic Classification of Digital Images of Archaeological Arrowheads
Currently there exist several databases composed of hundreds or thousands of digital images of arrowheads made by different ancient ethnic groups around the world. Extracting information or comparing and classifying the elements of these databases in an efficient and automated way, even without the need of arrowhead’s metadata, would be of great help in carrying out a comprehensive study on this archaeological subject. This work deals with this problem by developing an image processing computational algorithm that performs the automatic classification of the arrowheads. Particularly, the algorithm was used in the study of a set of Mexican’s arrowheads coming from the Michoacán state; these arrowheads were classified and compared with other databases consisting of hundreds of arrowheads from North America. The classifier uses these features: eccentricity of the arrowhead, blade kind (excurvate-incurvate) and base kind (concave-curvate, straight, with handle, and concave with handle). Except for the eccentricity, the other features were obtained by the curvature scale space (CSS) method computed on the arrowhead contour. An important aspect of the algorithm is its robustness with respect to the image, as it performs satisfactorily even with images of medium quality and the only general requirement is to have a uniform background.
Cite this Record
Automatic Classification of Digital Images of Archaeological Arrowheads. Fernando Castillo Flores, Francisco Javier García Ugalde, José Luis Punzo Díaz, Alfonso Gastelum-Strozzi, Dante Bernardo Martinez Vazquez. Presented at The 82nd Annual Meeting of the Society for American Archaeology, Washington, DC. 2018 ( tDAR id: 444592)
This Resource is Part of the Following Collections
min long: -108.853; min lat: 18.771 ; max long: -102.788; max lat: 25.76 ;
Abstract Id(s): 21464