# NAT-saliency

**Repository Path**: mirrors_NVlabs/NAT-saliency

## Basic Information

- **Project Name**: NAT-saliency
- **Description**: Noise Aware Training (NAT) for visual saliency
- **Primary Language**: Unknown
- **License**: Not specified
- **Default Branch**: main
- **Homepage**: None
- **GVP Project**: No

## Statistics

- **Stars**: 0
- **Forks**: 0
- **Created**: 2021-11-23
- **Last Updated**: 2025-09-20

## Categories & Tags

**Categories**: Uncategorized

**Tags**: None

## README

# Noise-Aware Video Saliency Prediction (BMVC 2021)

This is the repository for the code and the ForGED dataset accompanying our paper on the noise-aware training (NAT) paradigm for video saliency prediction published at BMVC 2021 ([arXiv link](https://arxiv.org/pdf/2104.08038.pdf)).

We plan to release a sample implementation of the proposed NAT loss function as well as the ForGED saliency dataset soon.