NPF-200
收藏资源简介:
NPF-200是首个大规模多模态非写实视频眼动数据集,由华南理工大学等机构创建。该数据集包含200个精心挑选的视频,超过70,000帧,涵盖了卡通和游戏视频,每个视频都配有音频,以辅助视觉注意力的检测。数据集的创建过程涉及使用Tobii眼动追踪器收集20名观察者的眼动数据,并将离散的眼动点转换为连续的注意力图。NPF-200数据集的应用领域包括媒体制作、艺术设计和游戏用户体验的增强,旨在解决非写实视频中视觉注意力的理解和预测问题。
NPF-200 is the first large-scale multimodal non-photorealistic video eye-tracking dataset, developed by South China University of Technology and other institutions. This dataset contains 200 carefully selected videos with over 70,000 frames in total, covering both cartoon and game videos, and each video is equipped with audio to aid visual attention detection. The development of the NPF-200 dataset involved collecting eye-tracking data from 20 observers using Tobii eye trackers, and transforming discrete eye fixation points into continuous attention maps. The application scenarios of the NPF-200 dataset include media production, art design, and enhancement of game user experience, aiming to address the issues of understanding and predicting visual attention in non-photorealistic videos.




