Using CLIP to extract and analyze images of the family in 3,000 Dutch-language children's books, 1800-1940.

This paper discusses the first step in a larger research project that uses computational techniques to analyze the multimodal representation of the family in the ~3,000 digitized Dutch-language children’s books (1800-1940) of the DBNL. We present a method that applies the multimodal machine learning (ML) model CLIP to classify images of families from the total 38,363 images in this corpus.