Look Into Past: A Different Perspective to Explore Cultural Heritage Datasets

Newspapers contain a wealth of historical information in the form of articles and illustrations. Library and cultural heritage institutions have been digitizing their collections for decades to enable web-based access and retrieval of information. Since digitized collections only include page-level metadata, it is challenging to find information about a specific context. In order to address this challenge, article-level information was segmented based on layout analysis techniques. Afterward, an image classification algorithm is used to filter articles that contain building images. An image's caption can provide information about its location. To validate the location information, a textual query is used to search Google for recent images of a historical location. By matching image features, irrelevant results will be filtered out. The historic images were then rephotographed, which is a notoriously challenging and complex process. These results can also be used by researchers in the humanities and cultural heritage to investigate spatiotemporal changes within a region or place.