# Dynamic Deep Multi-modal Fusion for Image Privacy Prediction

arXiv: https://arxiv.org/abs/1902.10796 · HTML: https://mlsanity.com/paper/1902.10796/

## Key facts

- Cited by 7 papers in the map (data through 14 Oct 2026).
- First posted on arXiv 27 Feb 2019.

## Built on by (most cited first, 7 of 7)

- [Evaluation of Human Visual Privacy Protection: A Three-Dimensional Framework and Benchmark Dataset](https://mlsanity.com/paper/2507.13981/) (2025)
- [DRAG: Dynamic Region-Aware GCN for Privacy-Leaking Image Detection](https://mlsanity.com/paper/2203.09121/) (2022)
- [Content-based Graph Privacy Advisor](https://mlsanity.com/paper/2210.11169/) (2022)
- [Image-guided topic modeling for interpretable privacy classification](https://mlsanity.com/paper/2409.18674/) (2024)
- [Learning Privacy from Visual Entities](https://mlsanity.com/paper/2503.12464/) (2025)
- [On the Robustness of Vision-Language Models in Zero-shot Privacy Classification](https://mlsanity.com/paper/2510.09253/) (2025)
- [Exploring Multi-Modality Dynamics: Insights and Challenges in Multimodal Fusion for Biomedical Tasks](https://mlsanity.com/paper/2411.00725/) (2024)

## Abstract (arXiv)

With millions of images that are shared online on social networking sites, effective methods for image privacy prediction are highly needed. In this paper, we propose an approach for fusing object, scene context, and image tags modalities derived from convolutional neural networks for accurately predicting the privacy of images shared online. Specifically, our approach identifies the set of most competent modalities on the fly, according to each new target image whose privacy has to be predicted. The approach considers three stages to predict the privacy of a target image, wherein we first identify the neighborhood images that are visually similar and/or have similar sensitive content as the target image. Then, we estimate the competence of the modalities based on the neighborhood images. Finally, we fuse the decisions of the most competent modalities and predict the privacy label for the target image. Experimental results show that our approach predicts the sensitive (or private) content more accurately than the models trained on individual modalities (object, scene, and tags) and prior privacy prediction works. Also, our approach outperforms strong baselines, that train meta-classifiers to obtain an optimal combination of modalities.