E-MLB
收藏资源简介:
E-MLB数据集是中国科学技术大学创建的大规模事件去噪数据集,包含100个场景,每个场景有四种噪声级别,总序列数达到1200,是现有最大数据集的12倍。该数据集覆盖了从白天到夜晚的不同光照条件,通过使用中性密度滤镜模拟不同光照强度,确保了数据的真实性和多样性。此外,数据集还包括多种运动类型,如平移、旋转以及2D和3D的组合运动。E-MLB数据集旨在为事件相机的去噪研究提供高质量、多样化的数据支持,解决现有数据集在规模和噪声多样性上的不足。
The E-MLB dataset is a large-scale event denoising dataset developed by the University of Science and Technology of China. It includes 100 scenarios, each with four noise levels, and has a total of 1200 sequences, which is 12 times the scale of the largest existing dataset. The dataset covers diverse lighting conditions ranging from daytime to nighttime, where neutral density filters are utilized to simulate varying light intensities, ensuring the authenticity and diversity of the data. Additionally, the dataset encompasses various motion types, such as translation, rotation, and combined 2D and 3D motions. The E-MLB dataset aims to provide high-quality and diversified data support for event camera denoising research, addressing the shortcomings of existing datasets in terms of scale and noise diversity.




