# An analysis of the transfer learning of convolutional neural networks for artistic images

arXiv: https://arxiv.org/abs/2011.02727 · HTML: https://mlsanity.com/paper/2011.02727/

## Key facts

- Cited by 11 papers in the map (data through 14 Oct 2026).
- First posted on arXiv 5 Nov 2020.

## Built on by (most cited first, 11 of 11)

- [Understanding and Creating Art with AI: Review and Outlook](https://mlsanity.com/paper/2102.09109/) (2021)
- [How Well do Feature Visualizations Support Causal Understanding of CNN Activations?](https://mlsanity.com/paper/2106.12447/) (2021)
- [Deconfounded Representation Similarity for Comparison of Neural Networks](https://mlsanity.com/paper/2202.00095/) (2022)
- [Deep transfer learning for image classification: a survey](https://mlsanity.com/paper/2205.09904/) (2022)
- [SniffyArt: The Dataset of Smelling Persons](https://mlsanity.com/paper/2311.11888/) (2023)
- [Art Authentication with Vision Transformers](https://mlsanity.com/paper/2307.03039/) (2023)
- [Addressing materials' microstructure diversity using transfer learning](https://mlsanity.com/paper/2107.13841/) (2021)
- [Interpretations of Domain Adaptations via Layer Variational Analysis](https://mlsanity.com/paper/2302.01798/) (2023)
- [Smelly, dense, and spreaded: The Object Detection for Olfactory References (ODOR) dataset](https://mlsanity.com/paper/2507.08384/) (2025)
- [Gesture Classification in Artworks Using Contextual Image Features](https://mlsanity.com/paper/2412.03456/) (2024)
- [Vision-based Warning System for Maintenance Personnel on Short-Term Roadwork Site](https://mlsanity.com/paper/2210.01689/) (2022)

## Abstract (arXiv)

Transfer learning from huge natural image datasets, fine-tuning of deep neural networks and the use of the corresponding pre-trained networks have become de facto the core of art analysis applications. Nevertheless, the effects of transfer learning are still poorly understood. In this paper, we first use techniques for visualizing the network internal representations in order to provide clues to the understanding of what the network has learned on artistic images. Then, we provide a quantitative analysis of the changes introduced by the learning process thanks to metrics in both the feature and parameter spaces, as well as metrics computed on the set of maximal activation images. These analyses are performed on several variations of the transfer learning procedure. In particular, we observed that the network could specialize some pre-trained filters to the new image modality and also that higher layers tend to concentrate classes. Finally, we have shown that a double fine-tuning involving a medium-size artistic dataset can improve the classification on smaller datasets, even when the task changes.