# Visual Saliency Prediction Using a Mixture of Deep Neural Networks

arXiv: https://arxiv.org/abs/1702.00372 · HTML: https://mlsanity.com/paper/1702.00372/

## Key facts

- Cited by 7 papers in the map (data through 14 Oct 2026).
- Placed in Vision mixture-of-experts, one of 12 directions in Mixture-of-experts.
- First posted on arXiv 1 Feb 2017.

## Placed in

- Mixture-of-experts / [Vision mixture-of-experts](https://mlsanity.com/moe/vision-mixture-of-experts/)

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

- [Spatio-Temporal Self-Attention Network for Video Saliency Prediction](https://mlsanity.com/paper/2108.10696/) (2021)
- [Learning from Observer Gaze:Zero-Shot Attention Prediction Oriented by Human-Object Interaction Recognition](https://mlsanity.com/paper/2405.09931/) (2024)
- [Where, What, Why: Towards Explainable Driver Attention Prediction](https://mlsanity.com/paper/2506.23088/) (2025)
- [FSDAM: Few-Shot Driving Attention Modeling via Vision-Language Coupling](https://mlsanity.com/paper/2511.12708/) (2025)
- [Learning Gaussian Representation for Eye Fixation Prediction](https://mlsanity.com/paper/2403.14821/) (2024)
- [From Scene to Object: Text-Guided Dual-Gaze Prediction](https://mlsanity.com/paper/2604.20191/) (2026)
- [FocusDrive: Reasoning with Visual Focus for Autonomous Driving](https://mlsanity.com/paper/2609.33190/) (2026)

## Abstract (arXiv)

Visual saliency models have recently begun to incorporate deep learning to achieve predictive capacity much greater than previous unsupervised methods. However, most existing models predict saliency using local mechanisms limited to the receptive field of the network. We propose a model that incorporates global scene semantic information in addition to local information gathered by a convolutional neural network. Our model is formulated as a mixture of experts. Each expert network is trained to predict saliency for a set of closely related images. The final saliency map is computed as a weighted mixture of the expert networks' output, with weights determined by a separate gating network. This gating network is guided by global scene information to predict weights. The expert networks and the gating network are trained simultaneously in an end-to-end manner. We show that our mixture formulation leads to improvement in performance over an otherwise identical non-mixture model that does not incorporate global scene information.