VQA-ODV
收藏资源简介:
VQA-ODV数据集由北京航空航天大学电子信息工程学院创建,包含600个全向视频序列,旨在研究全向视频的视觉质量评估(VQA)。数据集不仅提供序列的主观质量评分,还包括观察者的头部运动(HM)和眼睛运动(EM)数据。VQA-ODV数据集内容多样,涵盖不同内容、时长和分辨率,包括压缩和地图投影的损伤。该数据集的应用领域主要集中在全向视频的视觉质量评估,特别是如何通过考虑人类行为(如HM和EM)来提高VQA的准确性。
The VQA-ODV dataset was created by the School of Electronic and Information Engineering, Beihang University. It consists of 600 omnidirectional video sequences, and is designed for research on visual quality assessment (VQA) of omnidirectional videos. In addition to providing subjective quality scores for the sequences, the dataset also includes head movement (HM) and eye movement (EM) data from observers. The VQA-ODV dataset features diverse content, covering various content types, durations and resolutions, and encompasses distortions induced by compression and map projection. The primary application scope of this dataset focuses on visual quality assessment for omnidirectional videos, particularly on how to improve the accuracy of VQA by considering human behaviors such as HM and EM.

- 1Bridge the Gap Between VQA and Human Behavior on Omnidirectional Video: A Large-Scale Dataset and a Deep Learning Model北京航空航天大学电子信息工程学院 · 2018年



