Benchmarking, Analysis, and Structural Estimation of Deblurring (BASED) dataset

本研究创建了名为Benchmarking, Analysis, and Structural Estimation of Deblurring (BASED)的数据集，由莫斯科国立大学使用光束分离器技术构建。该数据集包含135个视频，专注于捕捉因物体快速移动而非相机移动产生的模糊。创建过程中，通过精确的光学设置和后处理技术，确保了数据的高质量和适用性。BASED数据集主要用于评估和改进图像去模糊技术，特别是在无参考图像质量评估方面的应用，旨在提高去模糊算法的准确性和效率。

This study developed a dataset named Benchmarking, Analysis, and Structural Estimation of Deblurring (BASED), which was constructed by Lomonosov Moscow State University using beam splitter technology. This dataset contains 135 videos, focusing on capturing blur induced by rapid movement of objects rather than camera motion. During its development, precise optical configurations and post-processing techniques were adopted to guarantee the high quality and applicability of the dataset. The BASED dataset is primarily intended for evaluating and advancing image deblurring technologies, particularly for applications in no-reference image quality assessment, with the aim of improving the accuracy and efficiency of deblurring algorithms.