Visual shape-matching and classification of ceramics

Created on 18. 5. 2019

A graphical user interface to a Deposit database of digitized ceramic fragments created using the Laser Aided Profiler control application which implements automated ceramic shape matching and clustering as described in Demján et al. 2022.

It has three basic functions: calculate (dis)similarity between samples; perform automatic clustering; and, allow users to rearrange clusters freely until they represent a valid classification. Furthermore, CeraMatch quantifies the degree of similarity of each fragment to a certain class, allowing for a probabilistic approach to classification, in which each vessel profile has quantifiable probabilities of belonging to particular classes or sub-classes. The application uses a hierarchical cluster analysis algorithm to assign each fragment into a cluster of similar specimens, a prerequisite for classification.

For an in-depth description of the concepts behind CeraMatch see:

DEMJÁN, P. - PAVÚK, P. - ROOSEVELT, CH. H. 2022: Laser-Aided Profile Measurement and Cluster Analysis of Ceramic Shapes. Journal of Field Archaeology 47 (8). https://doi.org/10.1080/00934690.2022.2128549

For a Windows installer see:

https://github.com/demjanp/ceramatch/releases/latest

To install the latest version of Ceramatch as a Python module use:

pip install ceramatch

To start the GUI execute:

bin\start_cm.py

Demján, P., Pavúk, P., Kaner, T., Bobik, J., & Roosevelt, C. H. (2021, October 15). Sample Data - Middle- to Late Bronze Age pottery from Kaymakçı. https://doi.org/10.17605/OSF.IO/UX8VD

You can create digital ceramics drawings usable in CeraMatch using the Laser Aided Profiler Control Application by digitizing pottery fragments directly using the LAP device, or by digitizing drawings in PDF or raster image format (see this video guide).

Peter Demján (peter.demjan@gmail.com)

Institute of Archaeology of the Czech Academy of Sciences, Prague, v.v.i.

Development of this software was supported by OP RDE, MEYS, under the project "Ultra-trace isotope research in social and environmental studies using accelerator mass spectrometry", Reg. No. CZ.02.1.01/0.0/0.0/16_019/0000728.

This software uses the following open source packages:

• cryptography
• Deposit GUI
• Deposit
• Graphviz
• lap_data
• natsort
• NetworKit
• NetworkX
• NumPy
• OpenCV on Wheels
• OpenCV
• openpyxl
• Pillow
• Psycopg
• PyGraphviz
• PyShp
• PySide2
• Qt
• scikit-image
• scikit-learn
• SciPy
• Shapely
• svgelements
• Unidecode
• validators

This code is licensed under the GNU GENERAL PUBLIC LICENSE - see the LICENSE file for details