Olfaction, often overlooked in cultural heritage studies, holds profound significance in shaping human experiences and identities. Examining historical depictions of olfactory scenes can offer valuable insights into the role of smells in history. To address the challenge of domain shift between natural images and artistic representations of fragrant spaces, we introduce novel scene-centric artistic datasets, RASD and WASD, comprising 704 and 3691 artworks respectively, constructed by querying two cultural heritage data sources, Rijksmuseum and Wikidata, and using the search terms as supervision signals. Additionally, Fragrant-Places dataset is created by manually reviewing all artworks of the ODOR dataset, consisting of 228 artworks. Furthermore, this complete set of ArtPlaces data is splited into ArtPlace_Train and ArtPlaces-test. ArtPlaces-test includes Fragrant-Places as well as a manually corrected portion of RASD and WASD. ArtPlace_Train contains the remaining sections of RASD and WASD. We show that a transfer-learning approach using weakly labeled training data can remarkably improve the classification of fragrant spaces and, more generally, artistic scene depictions. The models are evaluated on the two manually corrected test splits, Fragrant-Places and ArtPlaces-test. This work lays a foundation for further exploration of fragrant spaces recognition and artistic scene classification.