Researchers from Lingnan University and Shandong University developed an AI system to identify ancient Chinese plant seeds [1].
The technology addresses a critical specialist bottleneck in archaeology by automating the classification of botanical remains. This allows researchers to more efficiently study the evolution of human civilization and agricultural practices over millennia [1], [2].
The system was trained using a dataset of seeds spanning 5,000 years [1], [2]. Phys.org said the tool achieves a classification accuracy of more than 90% [1].
Botanical identification in archaeology typically requires rare expertise and significant time. The new system emulates the identification process of archaeobotanical experts to speed up the analysis of site finds [1].
By reducing the reliance on a small number of human specialists, the universities aim to accelerate the processing of archaeological data. The collaboration between the two institutions focuses on bridging the gap between computer science and historical research [1], [2].
“The system emulates the identification process of archaeobotanical experts”
The integration of AI into archaeobotany shifts the role of the specialist from manual identification to high-level data interpretation. By automating the classification of seeds, researchers can process larger volumes of material from diverse sites, potentially revealing new patterns in early crop domestication and climate adaptation that were previously too labor-intensive to map.
