EAA2020: Abstract
Abstract is part of session #464:
Title & Content
Title:
Machine learning approaches for optimized semi-automatic analysis of ancient grains
Content:
The archaeological science of today has reached beyond its initial purpose of collecting artifacts, a more prevalent anthropological dimension. It is not only sufficient to extract the valuable items from the excavation site and to preserve them properly, but seemingly unimportant elements from the premises of an archaeological site have become a golden source of information for researchers especially following the advent of molecular biology, genetics on the one hand and machine learning and artificial intelligence on the other.
The field of bioarchaeology studies not only the human and animal remains from the burial ground to detect genetically relevant correlations or clustering but also tackles the analysis of substances, microbiota and vegetal items that are also encountered on an archaeological site. Items that seemed worthless to archaeologists decades ago, now offer a plethora of information regarding more complex anthropological features of the ancient populations such as their diets, their health status, and their customs.
The present study proposes the employment of several computer vision and machine learning techniques that will allow researchers to obtain reliable meta-information on the cereal grains recovered from the Capidava fortress (Constanța County, Romania). The method enables computer vision techniques enriched with machine learning enhancements to automatically segment the images of grains, detect their contour as accurately as possible, analyze the textural feature of the segmented regions, cluster the regions based on the joint contour and textural priorly extracted features and then use the obtained clusters for an optimized semi-automated labeling of archaeological cereal grains.
Acknowledgments
The present work has received financial support through the project: Entrepreneurship for innovation through doctoral and postdoctoral research, POCU/360/6/13/123886 co-financed by the European Social Fund, through the Operational Program for Human Capital 2014- 2010.
The field of bioarchaeology studies not only the human and animal remains from the burial ground to detect genetically relevant correlations or clustering but also tackles the analysis of substances, microbiota and vegetal items that are also encountered on an archaeological site. Items that seemed worthless to archaeologists decades ago, now offer a plethora of information regarding more complex anthropological features of the ancient populations such as their diets, their health status, and their customs.
The present study proposes the employment of several computer vision and machine learning techniques that will allow researchers to obtain reliable meta-information on the cereal grains recovered from the Capidava fortress (Constanța County, Romania). The method enables computer vision techniques enriched with machine learning enhancements to automatically segment the images of grains, detect their contour as accurately as possible, analyze the textural feature of the segmented regions, cluster the regions based on the joint contour and textural priorly extracted features and then use the obtained clusters for an optimized semi-automated labeling of archaeological cereal grains.
Acknowledgments
The present work has received financial support through the project: Entrepreneurship for innovation through doctoral and postdoctoral research, POCU/360/6/13/123886 co-financed by the European Social Fund, through the Operational Program for Human Capital 2014- 2010.
Keywords:
machine learning, computer vision technique, archaeological cereal grains
Downloads:
authors
Main authors:
Cristina Mircea1,2
Co-author:
Ioan Gabriel Mircea3
Tiberiu Potârniche4
Beatrice Kelemen1,2
Tiberiu Potârniche4
Beatrice Kelemen1,2
Affiliations:
1 Department of Molecular Biology and Biotechnology, Faculty of Biology and Geology, Babes-Bolyai University, Cluj-Napoca, Romania
2 Molecular Biology Center, Interdisciplinary Research Institute on Bio Nano Sciences, Babes-Bolyai University, Cluj-Napoca, Romania
3 Department of Computer Science, Faculty of Mathematics and Informatics, Babes-Bolyai University, Cluj-Napoca, Romania
4 Museum of National History and Archeology, Constanța, Romania
2 Molecular Biology Center, Interdisciplinary Research Institute on Bio Nano Sciences, Babes-Bolyai University, Cluj-Napoca, Romania
3 Department of Computer Science, Faculty of Mathematics and Informatics, Babes-Bolyai University, Cluj-Napoca, Romania
4 Museum of National History and Archeology, Constanța, Romania