CeraMatch

0562392 - ARÚ 2023 RIV CZ eng L4 - Software
Demján, Peter
CeraMatch.
Internal code: CeraMatch 1.0.4 ; 2022
Technical parameters: 269 KB
Economic parameters: Software na vizuálnu klasifikáciu keramických fragmentov.
R&D Projects: GA MŠMT EF16_019/0000728
Keywords : computational ceramic classification * unsupervised machine-learning * automated shape matching
OECD category: Archaeology
https://github.com/demjanp/CeraMatch

A graphical user interface to a Deposit (https://github.com/demjanp/deposit) database of digitized ceramic fragments created using the Laser Aided Profiler control application (https://www.laseraidedprofiler.com/) 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. 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.
Permanent Link: https://hdl.handle.net/11104/0334814