SBI数据集
收藏资源简介:
SBI数据集是由国家研究委员会的高性能计算和网络研究所创建的,用于场景背景初始化的基准数据集。该数据集包含7个序列,来源于公开可用的数据集,如COST 211、ATON和PBI。数据集的创建过程包括手动选择或拼接无前景物体的序列帧作为真实背景。SBI数据集主要应用于视频监控和计算摄影等领域,旨在通过提供一个公共的真实背景模型,促进不同背景初始化方法的公平比较和性能评估。
The SBI dataset was created by the Institute of High Performance Computing and Networking of the National Research Council as a benchmark dataset for scene background initialization. It contains 7 sequences sourced from publicly available datasets such as COST 211, ATON, and PBI. The dataset creation process involves manually selecting or stitching sequence frames free of foreground objects as ground-truth backgrounds. The SBI dataset is primarily applied in fields including video surveillance and computational photography, aiming to facilitate the fair comparison and performance evaluation of different background initialization methods by providing a public ground-truth background model.




