Novel Artistic Scene-Centric Datasets for Effective Transfer Learning in Fragrant Spaces
Fuente:
arXiv
Saved in:
| Main Authors: | , , , , |
|---|---|
| Format: | Preprint |
| Published: |
2024
|
| Subjects: | |
| Online Access: | |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866913432871108608 |
|---|---|
| author | Liu, Shumei Huang, Haiting Zinnen, Mathias Maier, Andreas Christlein, Vincent |
| author_facet | Liu, Shumei Huang, Haiting Zinnen, Mathias Maier, Andreas Christlein, Vincent |
| contents | 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. 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. We fine-tune Places365-pre-trained models by querying two cultural heritage data sources and using the search terms as supervision signal. The models are evaluated on two manually corrected test splits. This work lays a foundation for further exploration of fragrant spaces recognition and artistic scene classification. All images and labels are released as the ArtPlaces dataset at https://zenodo.org/doi/10.5281/zenodo.11584328. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2407_11701 |
| institution | arXiv |
| publishDate | 2024 |
| record_format | arxiv |
| spellingShingle | Novel Artistic Scene-Centric Datasets for Effective Transfer Learning in Fragrant Spaces Liu, Shumei Huang, Haiting Zinnen, Mathias Maier, Andreas Christlein, Vincent Computer Vision and Pattern Recognition I.4.8; J.5 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. 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. We fine-tune Places365-pre-trained models by querying two cultural heritage data sources and using the search terms as supervision signal. The models are evaluated on two manually corrected test splits. This work lays a foundation for further exploration of fragrant spaces recognition and artistic scene classification. All images and labels are released as the ArtPlaces dataset at https://zenodo.org/doi/10.5281/zenodo.11584328. |
| title | Novel Artistic Scene-Centric Datasets for Effective Transfer Learning in Fragrant Spaces |
| topic | Computer Vision and Pattern Recognition I.4.8; J.5 |
| url | https://arxiv.org/abs/2407.11701 |