Examining the Merits of 3D Sliced Segmental Extraction: A Case Study of Prehistoric Japanese Ceramics

This article has been contributed by James Frances Loftus III. Information about the author is included at the end of the post.

Introduction

In recent years, the seemingly endless merits of utilizing 3D technologies in the digital preservation and online showcasing of archaeological materials has been increasingly showcased, with an influx of presentation sessions in major archaeological conferences dedicated solely to the dissemination of knowledge regarding the use of 3D technology (such as session #520 in the 2021 EAA annual meeting). With the onset of COVID19 and the subsequent upheaval of governmental institutions such as museums due to stay/shelter-at-home orders, the necessity of digitizing museum and institutional holdings of prehistoric materials became astonishingly apparent; with researchers unable to access necessary samples, and the scholarship facing a time of halting research progress. This was very much the situation in the past year within Japanese archaeology as well, with many scholars (especially those working with samples from distant prefectures) unable to collect primary data. Despite museums and housing institutions within Japan increasingly digitizing collections through 3D methods, still more needs to be done to examine the merits of utilizing 3D methods for research purposes on prehistoric Japanese samples, especially ceramics.

As such, this short piece will outline, in a manner also accessible to those not currently working with 3D materials, just one of many methods of using openly-accessible 3D databases and primary scanned data in the visualization of differential curvature of ceramic vessels, showcasing the usefulness of 3D morphometric mapping (3DMM) in extracting data for use in original research.

Methods

3D methods of quantifying data on ceramics is a particularly difficult task, as traditional morphometric analysis on 3D data often utilizes landmark-based methods, such as in the study of stone tools (Archer & Presnyakova, 2019).  However, such landmarks are difficult to utilize in ceramic studies due to a lack of homologous points to attach landmarks to, as well as the extremely curved nature of ceramic vessels (Wang & Marwick, 2020). As such, many modern studies rely on 2D methods of quantification, such as outline-based geometric morphometrics. However, 2D outlines only represent a very small degree of the total curvature of pottery vessels, which may retain certain traces of individual style left on the surface of pots during the production process, as seen through micro-variations in curvature (Gandon et al. 2020). As such, an alternative 3D method of visualizing the varying curvature of ceramics, despite being an underutilized line of research, is a particularly important and necessary step in the utilization of 3D methods in archaeology and material cultural studies in general.

Although this short study utilizes primary data, any open-access and downloadable 3D models can be utilized in the visualization of ceramic curvature, so long as the model is a recognizable file type in the CloudCompare software.* 3D models in this study were extracted using an Artec Space Spider 3D scanner (Fig.1), with scan orientation conducted within Artec Studio 15 (Arctec3D, Luxembourg) (Fig.2).

Figure 1. 3D Data Acquisition utilizing the Artec Space Spider.

Figure 2. Scanning and alignment of multiple individual scans to create a 3D mesh model.

The 3D visualization method used here is termed Sliced Segment Extraction (henceforth “SSE”). SSE takes the 3D mesh models of ceramic vessels and extracts point clouds for use in relative height measurements through the projection of color gradient data onto each point within the cloud, utilizing the CloudCompare software (CloudCompare 2.12, 2021) (Fig. 3-1). Following this, the cloud is “sliced” down the center and “unrolled” into a semi-3D state, from which it is possible to visually assess previously minute details in the curvature of vessels (Fig. 3-2). This semi-3D state is then projected into a 2D photo, showing extremely finely detailed inward-outward curvature on the surface of ceramics (Fig. 3-3).

Figure 3. Sliced Segmental Extraction (SSE) Workflow, inward-outward curvature is expressed through a color gradient (outward curvature being that of light blue, blue, green).

It is from this 2D photo, extracted from multiple steps of 3D visual analysis that potentials for incorporating 3D data into extracting potential individual style during ceramic production in prehistoric contexts is possible. A dedicated guide to SSE extraction is planned for a future study, as for now this study is simply dedicated to showcasing the usefulness of SSE in action.

Materials

This study utilizes prehistoric Japanese ceramics from the early Yayoi period, widely understood as the transitional period from mixed hunter-gatherer and low-yield agriculture to that of intensifying wet-rice agriculture in the Japanese archipelago (Nakayama, 2010; Yane, 1984). Within this period, the northern Kyushu region (北部九州 hokubu kyūshū) in particular is regarded as the initial region of contact with agriculturalists from the Korean peninsula, who migrated to the Japanese archipelago, spreading both agricultural technology as well as new pottery shape types among the Jomon peoples, who already had been producing widely variable pottery types for more than 10,000 years (Tanaka, 2011). In particular, the Fukuoka plain (福岡平野 fukuoka heiya) subregion (Fig.4) was seemingly a central pillar to the synthetization of Jomon and migrant pottery types, especially that of mortuary pottery, into the hybridized “Yayoi style”, which then was disseminated across western Japan in the 5^thc BC (Miyamoto, 2016, 2017).

Figure 4. Left: Map illustrating the research area within East Asia. Right: The greater Kyushu island region and chosen region of study, the northern Kyushu and associated sub-regions.

Mortuary pottery of this period, known as “tsubo” globular pots (壺型土器 tsubogata doki), henceforth “tsubo”, underwent several waves of morphological change (Loftus, 2021), and may have been increasingly standardized within the Fukuoka plain as cultural traditions were synthesized into the Yayoi culture. The Nakaterao site (中・寺尾遺跡 naka・terao iseki) is one of many early Yayoi mortuary sites within the Fukuoka plain region, consisting of a mix of pit burials associated with late Jomon period traditions, and newly developed jar burials (Ono/Onojo City Board of Education, 1971, 1977, 1999). Remaining burial furniture during this early period often only consists of the tsubo mortuary vessel, as other possible goods made from natural materials may have deteriorated in the long span of 2,000 years. Tsubo vessels recovered from the Nakaterao site seem to be much more varied than that of other vessels excavated from other major Fukuoka plain sites, as expressed through comparative CV (coefficient of variance) analysis on principal component scores extracted from 2D outline-based morphometrics (Fig. 5).

Figure 5. Comparative CV analysis (based on standardized PC scores) of early Yayoi period, Fukuoka plain mortuary pottery. Sites with 10+ samples are bolded, showing the Nakaterao site to be one of the most variable in ceramic morphology (*A: PC2 = 23% variance, lower body/base; *B: PC1 = 31% variance, lip/neck).

Within sites with more than ten samples, the Nakaterao site shows exceptionally varied 2D shape of pottery. As such, the pottery from this site is especially useful in understanding exactly which region of the vessel is the most variable in curvature through the use of 3DMM and SSE. As this piece is only an overview of the merits of these 3D methods, only two representative vessels are chosen for 3D visualization at this time (samples A & B, Fig.6).

Figure 6. 3D digitized samples utilized in this study (Nakaterao site, Onojo city).

Results

As visualized below (Fig. 7), intra-vessel inward-outward variance in curvature is extremely variable within the two example vessels provided. In particular, the highest variable range of curvature exists in two main regions, the lip/neck and middle body areas of vessels. These results match well with results from 2D outline-based morphometrics (Fig.5), which show that shape variance is highest within these two regions in particular. Different from that of 2D results however, 3DSSE is able to pinpoint exactly in what ways these regions are variable.

In the lip/neck region, outward curvature seems to be relatively confined to one area of the region in sample A, and the outward curvature of the lip and neck both align vertically, suggesting that this region was mostly likely attached to the body of the vessel in one piece, or intensively smoothed in a similar manner, rather than in multiple stages which would shift in directionality. However, sample B shows that rather than one area of outward curvature on the lip/neck, two areas of curvature that create an oblong shape exist. Furthermore, the outward curvature of the lip is only vertically aligned with the lip in one area, suggesting that the smoothing process was unable to match that of the neck, or multiple stages of clay building while rotating the vessel was necessary.

In the body region, sample A shows clear circular regions of outward curvature in two areas, which would be found on opposite sides of the middle of the body, suggesting that instead of a perfectly rounded globular body, of which was the “normal” style at the time, this individual was creating vessels with a slightly oblong or egg-shaped form. Sample B on the other hand shows curvature contained to one region only, directly aligned below the curvature of the lip/neck as well as the base, this may shed light on the motor skills used by this individual, and/or the sitting position when coiling the vessel, with the body slightly curved outward toward the individual as they smoothed the surface of the vessel.

Figure 7. Nakaterao samples A & B, with SSE extracted. Curvature is expressed through color gradients from outward – inward curvature (Outward: light blue, blue, green, yellow; Inward: orange, red, white).

Discussion & Conclusions

The results above highlight how 3DMM is able to accurately deduce complexities in the form of ceramics, even those of prehistoric samples. In particular, samples from the Nakaterao site show high curvature variance both in the attachment process of the lip/neck region, and the general roundness of the body shape, with most examples showing more of an oblong curvature to the vessels. Previously, without direct access to such samples, visualizing differential curvature in such a detailed way was not possible through the use of 2D line drawings in a site report. As many disciplines move towards digital methods (such as in the recent boom in digital humanities), further accelerated by restricted access to samples due to COVID19; it is this author’s belief that methods of taking open-access 3D models beyond simply showcasing them in online exhibitions, and into more mainstream academic analysis will eventually become more commonplace in coming years. As such, increasing the abundance of accessible methods of both visualization (such as with 3DMM), and quantification will become increasingly necessary. It is hoped that through a short and approachable study such as this, that the merits of combining currently developed 2D methods with underutilized 3D workflows such as this will become clear to not only ceramic experts, but those working in a multitude of material cultural studies both within Japan and abroad.

---

Notes

* A list of compatible files can be found on the CloudCompare wiki, see: https://tinyurl.com/4c2mmy3x

References

Archer, W., & Presnyakova, D. 2019. Considerations in the application of 3DGM to stone artifacts with a focus on orientation error in bifaces. In Seguchi, N., Dudzik, B. 2019. 3D Data Acquisition for Bioarchaeology, Forensic Anthropology, and Archaeology, pp. 161-173. Cambridge: Academic Press. https://doi.org/10.1016/B978-0-12-815309-3.00008-5.

Wang, L. & Marwick, B. 2020. Standardization of ceramic shape: A case study of Iron Age pottery from northeastern Taiwan. Journal of Archaeological Science: Reports 33. https://doi.org/10.1016/j.jasrep.2020.102554.

Gandon E. et. al. 2020. Traditional craftspeople are not copycats: Potter idiosyncrasies in vessel morphogenesis. PLOS ONE 15(9): e0239362. https://doi.org/10.1371/journal.pone.0239362.

CloudCompare (version 2.12) [GPL software]. 2021. Retrieved from http://www.cloudcompare.org/

Nakayama, S. 2010. Botanical Archaeology and Origin of Agriculture in Japan. Tokyo: Doushisha. (In Japanese)

Yane, Y. 1984. Jomon dokki kara yayoi dokki e [From Jomon pottery to Yayoi pottery]. Tokyo. (In Japanese)

Tanaka, Y. 2011. Yayoi jidai shūraku ato no kenkyū [Yayoi Settlement Ruin Research]. Tokyo: Shinsensha. (In Japanese)

Miyamoto, K. 2016. Archaeological Explanation for the Diffusion Theory of the Japonic and Koreanic Languages. Japanese Journal of Archaeology (4).

Miyamoto, K. 2017. Higashi ajia no soki inasaku to yayoi no kigen [Early agriculture in northeast Asia and the origin of Yayoi]. Doseisha: Tokyo. (In Japanese)

Loftus, J.F. III. Micro-Regional Complexities in Yayoi Pottery Form as seen through Morphometric Statistical Analysis. Japanese Journal of Archaeology. December, 2021 (expected).

Ono City Board of Education. 1971. Naka Terao iseki [Excavation Report of the Naka Terao Site]. Ono: Ono City Board of Education. 大野町教育委員会 1971 『中・寺尾遺跡』大野町教育委員会

Onojo City Board of Education. 1977. Naka Terao iseki [Excavation Report of the Naka Terao Site]. Onojo: Onojo City Board of Education. 大野町教育委員会 1977 『中・寺尾遺跡』大野町教育委員会

Onojo City Board of Education. 1999. Naka Terao iseki III [Excavation Report of the Naka Terao Site]. Onojo: Onojo City Board of Education. 大野城市教育委員会 1999 『大野城市文化財調査報告書54：中・寺尾遺跡III』大野城市教育委員会

About the Author

James Loftus is a PhD candidate at Kyushu University, and a Research Fellow for the Japan Society for the Promotion of Science. His current project focuses on the evolution of social learning methods and human cognition during the agricultural transitional period (late Jomon – early Yayoi) throughout the Japanese Archipelago. This is done by utilizing geometric morphometric analysis on a variety of earthenwares. You can find his current work on ResearchGate (JamesFLoftusIII), or daily updates regarding archaeology, life in Japan, and being Hafu in Japan on his Twitter (@JamesFLoftusIII).

Cover Image

A 3D digitized early Yayoi period mortuary vessel from the Nakaterao site; housed at the Onojo Cocoro-no-furusato-kan City Museum, Onojo city, Fukuoka, Japan.