UBC3Deye
收藏资源简介:
UBC3Deye数据集是由不列颠哥伦比亚大学电气工程系和ICICS创建的大型公开可用眼睛追踪数据集,包含61个立体3D视频及其对应的2D版本。数据集通过24名参与者在自由观看测试中收集眼睛追踪数据,旨在为研究社区提供一个用于验证不同3D视觉注意力模型的基准。数据集涵盖了广泛的强度、运动、深度和纹理密度,确保了数据的多样性和代表性。此外,数据集还包括一个在线基准系统,用于验证和比较现有的2D和3D视觉注意力模型,并支持新模型的添加。该数据集的应用领域主要集中在3D视频内容的视觉注意力预测和模型验证,旨在解决3D视觉注意力模型的准确性和有效性问题。
The UBC3Deye dataset is a large publicly available eye-tracking dataset developed by the Department of Electrical Engineering and ICICS at the University of British Columbia. It includes 61 stereoscopic 3D videos along with their corresponding 2D versions. Eye-tracking data for the dataset was collected from 24 participants during free-viewing tests, with the primary purpose of providing the research community with a benchmark to validate diverse 3D visual attention models. The dataset covers a broad spectrum of intensity, motion, depth, and texture density, ensuring the diversity and representativeness of the collected data. Furthermore, the dataset features an online benchmark system that allows for the verification and comparison of existing 2D and 3D visual attention models, and supports the integration of new models. The primary application areas of this dataset are focused on visual attention prediction and model validation for 3D video content, aiming to address the accuracy and effectiveness challenges of 3D visual attention models.

- 1Benchmark 3D eye-tracking dataset for visual saliency prediction on stereoscopic 3D video不列颠哥伦比亚大学 · 2018年



