Applying Deep Learning in Creative Re-creation of Changsha Kiln Cultural Relics

Abstract

Changsha Kiln, world-renowned for its rich under-glazed porcelain, is a famous export porcelain kiln in the Tang Dynasty, began to thrive during the middle and late periods of the Tang Dynasty (618–907 AD), but then declined during the Five Dynasty periods (907–960 AD). Since the beginning of the new century, with the country’s increasing emphasis on the protection and innovation of cultural heritage, Changsha Kiln has gradually moved towards revival. However, the lack of ceramic products design talents, difficult to blend traditional and modern styles, and lack of support for creative design are obstacles to revival. To this end, this paper proposes an open creative design platform for cultural relics re-creation in Changsha kiln. Three basic components built based on deep learning technology in this platform: cultural relics knowledge base, cultural relic image feature database, and search engine based on semantics and images. With this platform, provide cultural relics element retrieval and creative design services for the general public, cultural creative designer and SMEs, which can promote the integrated development of culture and technology, and promote the cultural industry to become a pillar industry of the national economy.