Publicación: Deep learning for the taphonomic study of bone surface modifications: Lion and jaguar tooth marks analysis through convolutional neural networks
No hay miniatura disponible
Fecha
2021-06-02
Autores
Editor/a
Director/a
Tutor/a
Coordinador/a
Prologuista
Revisor/a
Ilustrador/a
Derechos de acceso
info:eu-repo/semantics/openAccess
Título de la revista
ISSN de la revista
Título del volumen
Editor
Universidad Nacional de Educación a Distancia (España). Escuela Técnica Superior de Ingeniería Informática. Departamento de Inteligencia Artificial
Resumen
La adecuada y precisa identificación y comprensión de las alteraciones tafonómicas hechas por carnívoros, entre otras modificaciones en las superficies de los huesos, es de enorme importancia para entender el registro arqueológico y para reconstruir el comportamiento humano y su interacción con otras especies. Como parte crucial, pero difícil, de la Tafonomía, ha sido el centro de numerosas controversias debido a su subjetividad y a la dificultad para identificar e interpretar las diferentes alteraciones como resultado de los diferentes métodos de aproximación, que varían según el autor. En los últimos años, los algoritmos de aprendizaje automático han ganado importancia como resultado de la búsqueda de herramientas y métodos más objetivos para el estudio de las alteraciones. Hemos usado una muestra de 250 surcos de diente hechos por Panthera leo y Panthera onca para comparar la enorme precisión de los algoritmos de aprendizaje y la alcanzada por tafónomos expertos. Además, este método ha demostrado su enorme potencial y eficacia clasificando y diferenciando las marcas de dientes creadas por leones y las creadas por jaguares, con una precisión de un 85%, a pesar de que tradicionalmente los estudios prácticos las han considerado indistinguibles.
A proper and accurate identification and comprehension of taphonomic alterations made by carnivores, among other bone surface modifications (BSM), is of capital importance in order to understand the archaeological record and to reconstruct the human behaviour and the interaction with other species. As a crucial part of Taphonomy, this topic has been the focus of numerous controversies due to subjectivity and difficulty when identifying and interpreting different alterations as the result of the variety of approaches depending on the expert. In recent years, machine learning algorithms have gained importance because of the searching for more objective tools and methods for the study of BSM. Here, we have used a sample of 250 tooth scores made by Panthera leo and Panthera onca to compare the high accuracy of computer trained algorithms and the one reached by expert taphonomists. This method has demonstrated its high potential and effectiveness classifying and differentiating tooth marks created by lions from the ones created by jaguars, with an accuracy of an 85%, despite the fact it has been traditionally claimed by many practical studies that they are indistinguishable.
A proper and accurate identification and comprehension of taphonomic alterations made by carnivores, among other bone surface modifications (BSM), is of capital importance in order to understand the archaeological record and to reconstruct the human behaviour and the interaction with other species. As a crucial part of Taphonomy, this topic has been the focus of numerous controversies due to subjectivity and difficulty when identifying and interpreting different alterations as the result of the variety of approaches depending on the expert. In recent years, machine learning algorithms have gained importance because of the searching for more objective tools and methods for the study of BSM. Here, we have used a sample of 250 tooth scores made by Panthera leo and Panthera onca to compare the high accuracy of computer trained algorithms and the one reached by expert taphonomists. This method has demonstrated its high potential and effectiveness classifying and differentiating tooth marks created by lions from the ones created by jaguars, with an accuracy of an 85%, despite the fact it has been traditionally claimed by many practical studies that they are indistinguishable.
Descripción
Categorías UNESCO
Palabras clave
surco, modificaciones de la superficie ósea, aprendizaje profundo, aprendizaje automático, redes neuronales convolucionales, Tafonomía, tooth score, bone surface modifications, deep learning, machine learning, convolutional neural networks, Taphonomy
Citación
Centro
Facultades y escuelas::E.T.S. de Ingeniería Informática
Departamento
Inteligencia Artificial