Abstract
Information and communication technologies (ICT) have been at the centre of most innovations. With applications in science, technology, engineering, and mathematics fields, it has become prevalent in business, art, and humanities disciplines, among others, as well. Among the potential applications of ICT in social sciences and digital humanities, documentation and reconstruction of archaeological artefacts have garnered interest and resulted in several studies. This is because of the potential inherent in these artefacts for archaeological and historical studies. However, regarding pottery reassembly, challenges are experienced in implementing an optimal solution entailing high standards. Although existing studies attempted to solve these challenges, a high standard solution is still elusive. This article presents an approach to a machine-driven solution that intends to use computer vision and machine learning, whose potential is yet to be felt in pottery reassembly. This investigation, still at an early stage, has profound implications for future studies in pottery studies in general.
TopBackground
As archaeologists study the historic and prehistoric human activities through reconstructing discovered ancient artefacts to have a grounded understanding of the prehistoric culture, the ability to carry out these reconstructions accurately and within reasonable time become pertinent. In this technology era of virtual and augmented reality, among others, researchers have been applying the use of digital technologies to make work faster and output optimal. This has helped in making professional life easier, more enjoyable and more productive, leading to what is now known as ‘smart environment’. Therefore, applying technological means to solve critical issues in the archaeological field is desirable.
Hence, many studies have attempted to solve the reconstruction problem. For example, the study of Kampel and Sablatnig (2003) developed a system that can process both complete and broken vessels. This was achieved using two reconstruction strategies known as: “shape from silhouette based method for complete vessels and a profile based method for fragments” (Kampel & Sablatnig, 2003). While using these strategies have improved performance with an acceptable accuracy, it is nonetheless dependent on certain conditions being met, thus requiring further investigations that will improve accuracy.
Key Terms in this Chapter
Pottery Reassembly/Reconstruction: The act of restoring potsherds, physically, or digitally, so that it looks as its original condition as much as possible.
Computer Vision: The ability to make computers or machines to see, interpret, and do similar tasks that humans can do.
Machine Learning: The science of getting computers to learn and improve their learning over time in an accurate and automatic manner, by giving them data and information.
Pottery Documentation: The process of acquiring, classifying, processing, and recording pottery objects, physically or digitally, for preservation and research purposes.
Pottery Classification: A procedure or process of systematic sorting and grouping together of pottery or potsherd of high similarities and characteristics.
Cultural Heritage: An expression of the way of life of a people, inherited and passed on from generation to generation and preserved in the present for posterity. The way of life of a people include its customs, traditions, values, knowledge, belief, law, art, and anything that is acquired and passed on.
Pottery: Vessels made from soft clay and hardened by heat.
Potsherd: A broken piece of pottery, or earthenware generally, with archaeological value.
Image Acquisition: The process of capturing an unprocessed image from an object or scene by an optical device into a manageable form for processing and analysis purposes.