Principal Investigator(s): Lin Du, Asian Languages and Cultures
Art historians and cultural historians use visual analysis to find patterns in primary sources because visual patterns can uncover the lineage of images, providing insight into pictorial knowledge production. To meet this need, we designed an Intelligent Reverse Image Retrieval system for Digital Humanities. The primary aim of this project is to develop an open-source reverse image retrieval system that will search historical images based on their content in an automated way. In our project, we propose solving popular retrieval problems in the digital humanities, by using a photograph or illustration as a query to search for its target image based on its local features within a database. Currently, our model has a 70% accuracy rate, making it the first to use interdisciplinary computer vision and visual analysis to analyze historical images in Chinese studies.