Identification of porcelain ewers in Tang, Song, and Yuan dynasties by digital shape characterization

As an important carrier of history, ancient ceramic ware is a crucial part of the traditional Chinese culture and heritage. The porcelain ewer is a typical example of ancient Chinese ceramic ware. The shapes of the porcelain ewers from the Tang, Song, and Yuan dynasties reflect their artistic characteristics. Different shapes and characteristics reflect the cultures of different dynasties and their production processes and lifestyle. Shapes are an important feature for identifying ancient ceramics. The complex structure of a porcelain ewer makes it difficult to reflect the evolution of its overall shape with partial shape changes. In this study, digital extraction methods of shape characteristics for porcelain ewers were discussed using samples of porcelain-handled ewers from the Tang dynasty to the Song and Yuan periods. Spatial correlation weights of same-side pixels and spatial similarity weights of symmetric pixels in the digital image were calculated by affinity measurement combined with the simple linear iterative cluster normalized cut (SLIC-Ncut) algorithm to segment porcelain ewers. The shape data of the main parts were extracted for analysis and preliminary classification. The position of the spout, the position of the handle, and the centre of gravity obtained from the shape data are crucial parameters for dating its age. The conclusions obtained are consistent with the traditional identification results, which indicate that the proposed method is an effective and feasible computer-aided method for identifying porcelain ewers.