TIMo (TIMo (Time-of-Flight Indoor Monitoring))

TIMo（飞行时间室内监控）是一个红外和深度视频数据集，旨在用于异常检测和人员检测/人数统计。它具有超过 1,500 个用于异常检测的序列，总计超过 500,000 个单独的帧。对于人员检测，数据集包含超过 240 个序列。数据是使用 Microsoft Azure Kinect RGB-D 相机捕获的。此外，我们提供异常帧范围的注释，用于异常检测和边界框，以及用于人员检测的分割掩码。数据是从倾斜视图和自上而下的角度部分捕获的。

TIMo (Time-of-Flight Indoor Monitoring) is an infrared and depth video dataset designed for anomaly detection, person detection and crowd counting. It contains over 1,500 sequences dedicated to anomaly detection tasks, totaling more than 500,000 individual frames. For person detection, the dataset includes over 240 sequences. All data was captured using a Microsoft Azure Kinect RGB-D camera. Additionally, we provide annotations including anomaly frame ranges for anomaly detection, bounding boxes, and segmentation masks for person detection. The data was partially captured from both oblique viewpoints and top-down perspectives.

搜集汇总

背景与挑战

背景概述

TIMo数据集包含超过1,500个异常检测序列和240个人员检测序列，总计超过500,000帧，提供异常帧范围注释和边界框及分割掩码，适用于室内监控场景。