Researchers have developed a pipeline called "Lot Machine" to automatically extract structured lot-level metadata from historical auction catalogs. This system utilizes vision-language models (VLMs) and is designed to address the challenges of variable formatting and the need for machine-readable data in provenance research and art market studies. The pipeline was evaluated across different deployment modes, including commercial endpoints, institutional gateways, and local deployments, demonstrating its feasibility and effectiveness for large-scale automated analysis of cultural heritage data. AI
IMPACT Enables large-scale automated analysis of historical auction catalogs, aiding provenance research and art market studies.
RANK_REASON Research paper detailing a new method for data extraction. [lever_c_demoted from research: ic=1 ai=1.0]
AI-generated summary · Google Gemini · from 1 sources. How we write summaries →