Search papers, labs, and topics across Lattice.
The NTIRE 2026 challenge on video saliency prediction tasked participants with developing algorithms to predict saliency maps from a novel dataset of 2,000 diverse videos. Fixation data was collected via crowdsourced mouse tracking from over 5,000 assessors. The challenge culminated in the evaluation of 7 teams on a held-out set of 800 videos using standard saliency metrics, with all data released publicly.
A new open-license dataset of 2,000 diverse videos with corresponding mouse-tracking saliency data provides a valuable benchmark for advancing video saliency prediction models.
This paper presents an overview of the NTIRE 2026 Challenge on Video Saliency Prediction. The goal of the challenge participants was to develop automatic saliency map prediction methods for the provided video sequences. The novel dataset of 2,000 diverse videos with an open license was prepared for this challenge. The fixations and corresponding saliency maps were collected using crowdsourced mouse tracking and contain viewing data from over 5,000 assessors. Evaluation was performed on a subset of 800 test videos using generally accepted quality metrics. The challenge attracted over 20 teams making submissions, and 7 teams passed the final phase with code review. All data used in this challenge is made publicly available - https://github.com/msu-video-group/NTIRE26_Saliency_Prediction.