VASESKETCH: Automatic 3D Representation of Pottery from Paper Catalog Drawings

Banterle, Francesco; Itkin, Barak; Dellepiane, Matteo; Wolf, Lior; Callieri, Marco; Dershowitz, Nachum; Scopigno, Roberto

doi:10.1109/icdar.2017.117

Cited by 10 publications

(8 citation statements)

References 18 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…As the 3D analysis of ceramics covers a wide landscape, the full breadth of scholarship cannot be included here, but recent studies can generally be categorized into three groups: 1) reconstruction/ identification; 2) 3D to 2D quantification; and 3) fullvessel extraction. Studies that utilize novel 3D software to reconstruct ceramic vessels from single sherds, piece together a collection of sherds into a full vessel (Banterle et al 2017;Stamatopoulos & Anagnostopoulos 2017), or seek to identify types of pottery through 3D imaging fall into the first group of recent studies, i.e., those that focus primarily on reconstruction or identification. Studies that utilize 3D mesh or cloud data as a base, but then extract 2D outline or other 2D data (such as length measurements) fall into the second group of studies (recent examples include Göttlich et al 2021 andHarush et al 2020).…”

Section: D and 3d Morphometric Analysis Of Ceramicsmentioning

confidence: 99%

Visualizing Variability in Ceramic Micro-Curvature though Novel 3D Morphometric Mapping and Sliced Segmental Extraction

III¹,

Seguchi²

2022

Journal of Computer Applications in Archaeology

View full text Add to dashboard Cite

In recent years, two-dimensional (2D) outline-based morphometric analysis has begun to be the standard means of shape analysis in the study of ceramics. Despite the everincreasing repertoire of three-dimensional (3D) ceramic data available to researchers, due to challenges in utilizing landmark-based geometric morphometrics on ceramics, many contemporary studies are forced to rely on 2D outlines, which account for only a small degree of shape variance. Consequently, notions of intra-vessel variance cannot be thoroughly extrapolated, and as such, important themes such as morphogenesis during the production process have not been thoroughly examined. By contrast, the present project utilizes a method derived from high-resolution 3D scans combining "morphometric mapping" and "sliced segmental extraction" to visualize previously unrecognized variability in the micro-curvature of full vessels. This method does not require homologous landmarks, but rather, employs a multistage approach that allows for several stages of analysis that can be adjusted to each given research project, and thus has the potential to elucidate idiosyncratic variability during the morphogenesis process.

show abstract

Section: D and 3d Morphometric Analysis Of Ceramicsmentioning

confidence: 99%

Visualizing Variability in Ceramic Micro-Curvature though Novel 3D Morphometric Mapping and Sliced Segmental Extraction

III¹,

Seguchi²

2022

Journal of Computer Applications in Archaeology

View full text Add to dashboard Cite

show abstract

“…An automatic classification method for the digitization of pottery profile drawings is presented by Banterle et al (2017). In this work, a structured description of the main geometric features is extracted and then, a 3D representation of each class is generated based on the geometric features.…”

Section: Performed Studies From 2010 To 2019mentioning

confidence: 99%

“…With the advancement of data collection tools, the use of machine learning systems to classify is increasing. If the data is large enough, the system can train to automatically classify new data (Banterle et al, 2017;Funkhouser et al, 2011;Toler-Franklin et al, 2010).…”

Section: Data Acquisition and Preprocessingmentioning

confidence: 99%

“…The use of the cost function has also been investigated (Papaioannou et al, 2002;Rasheed & Nordin, 2015). It should be noted that, in recent years, attention in using a machine learning system for matching is increasing (Banterle et al, 2017;Rasheed & Nordin, 2018). The used reconstruction methods in the analyzed literature are shown in Figure 6: the most commonly used approaches are Meta-heuristic optimization and Similarity Analysis.…”

Section: Data Acquisition and Preprocessingmentioning

confidence: 99%

See 1 more Smart Citation

Review of computer-based methods for archaeological ceramic sherds reconstruction

Eslami

Angelo

Stefano

et al. 2020

Virtual archaeol. rev.

View full text Add to dashboard Cite

Potteries are the most numerous finds found in archaeological excavations; they are often used to get information about the history, economy, and art of a site. Archaeologists rarely find complete vases but, generally, damaged and in fragments, often mixed with other pottery groups. By using the traditional manual method, the analysis and reconstruction of sherds are performed by a skilled operator. Reviewed papers provided evidence that the traditional method is not reproducible, not repeatable, time-consuming and its results have great uncertainties. To overcome the aforementioned limits, in the last years, researchers have made efforts to develop computer-based methods for archaeological ceramic sherds analysis, aimed at their reconstruction. To contribute to this field of study, in this paper, a comprehensive analysis of the most important available publications until the end of 2019 is presented. This study, focused on pottery fragments only, is performed by collecting papers in English by the Scopus database using the following keywords: “computer methods in archaeology", "3D archaeology", "3D reconstruction", "automatic feature recognition and reconstruction", "restoration of pottery shape relics”. The list is completed by additional references found through the reading of selected papers. The 53 selected papers are divided into three periods of time. According to a detailed review of the performed studies, the key elements of each analyzed method are listed based on data acquisition tools, features extracted, classification processes, and matching techniques. Finally, to overcome the actual gaps some recommendations for future researches are proposed.Highlights:<ul><li>The traditional manual method for reassembling sherds is very time-consuming and costly; it also requires a great deal effort from skilled archaeologists in repetitive and routine activities.</li><li>Computer-based methods for archaeological ceramic sherds reconstruction can help archaeologists in the above-mentioned repetitive and routine activities.</li><li>In this paper, the state-of-the-art computer-based methods for archaeological ceramic sherds reconstruction are reviewed, and some recommendations for future researches are proposed.</li></ul>

show abstract

“…(Cf. (Banterle et al 2017). ) However, to avoid the associated computation overhead of 3D reconstruction, we propose a method to obtain the synthetic fracture surface (Fig 1(f)) directly from the 2D catalog sketch, and even on-the-fly during training.…”

Section: Introductionmentioning

confidence: 99%

Computational Visual Ceramicology: Matching Image Outlines to Catalog Sketches

Itkin

Wolf

Dershowitz

2021

AAAI

Self Cite

View full text Add to dashboard Cite

Field archeologists are called upon to identify potsherds, for which they rely on their professional experience and on reference works. We have developed a recognition method starting from images captured on site, which relies on the shape of the sherd's fracture outline. The method sets up a new target for deep-learning, integrating information from points along inner and outer surfaces to learn about shapes. Training the classifiers required tackling multiple challenges that arose on account of our working with real-world archeological data: paucity of labeled data; extreme imbalance between instances of different categories; and the need to avoid neglecting rare classes and to take note of minute distinguishing features of some classes. The scarcity of training data was overcome by using synthetically-produced virtual potsherds and by employing multiple data-augmentation techniques. A novel form of training loss allowed us to overcome classification problems caused by under-populated classes and inhomogeneous distribution of discriminative features.

show abstract

VASESKETCH: Automatic 3D Representation of Pottery from Paper Catalog Drawings

Cited by 10 publications

References 18 publications

Visualizing Variability in Ceramic Micro-Curvature though Novel 3D Morphometric Mapping and Sliced Segmental Extraction

Visualizing Variability in Ceramic Micro-Curvature though Novel 3D Morphometric Mapping and Sliced Segmental Extraction

Review of computer-based methods for archaeological ceramic sherds reconstruction

Computational Visual Ceramicology: Matching Image Outlines to Catalog Sketches

Contact Info

Product

Resources

About