TIMo—A Dataset for Indoor Building Monitoring with a Time-of-Flight Camera

Abstract

We present TIMo (Time-of-flight Indoor Monitoring), a dataset for video-based monitoring of indoor spaces captured using a time-of-flight (ToF) camera. The resulting depth videos feature people performing a set of different predefined actions, for which we provide detailed annotations. Person detection for people counting and anomaly detection are the two targeted applications. Most existing surveillance video datasets provide either grayscale or RGB videos. Depth information, on the other hand, is still a rarity in this class of datasets in spite of being popular and much more common in other research fields within computer vision. Our dataset addresses this gap in the landscape of surveillance video datasets. The recordings took place at two different locations with the ToF camera set up either in a top-down or a tilted perspective on the scene. Moreover, we provide experimental evaluation results from baseline algorithms.

A Review on Video Anomaly Detection Datasets

“…The Subway dataset [50] is collected from two scenes: an entrance and an exit of a subway. The Time-of-flight Indoor Monitoring (TIMO) dataset [51] was recently developed. It contains indoor spaces taken using a time-offlight (ToF) camera [52] from two different views: topdown or tilted perspective.…”

Time-of-Flight Depth Sensing for Automotive Safety and Smart Building Applications: The VIZTA Project

“…Due to the absence of any publicly available datasets of ToF camera images covering our target application areas of in- car cabin monitoring and smart building person and anomaly detection, we utilized our experimental setup environments described above in Section III in order to record two new ToF datasets. The datasets TICaM (in-car) and TIMo (Indoor monitoring) were published in [19] and [20] respectively, and were made publicly available for research purposes through our website https://vizta-tof.kl.dfki.de/.…”

“…Same as in the case of in-cabin monitoring, we recorded and annotated our own dataset, TIMo [20] for our indoor monitoring cases. The dataset consists of two sub-parts:…”

“…The detection networks for the studies on person detection in buildings [19], [35], [39] and in-cabin occupant detection [20] were implemented in Python using the PyTorch framework with the possibility of further network conversion to ONNX format with opset 11. The training was done on an Ubuntu 18.04 x64 PC with 4 Nvidia GTX 1080Ti Graphics Processing Units (GPUs).…”

“…Although our two introduced benchmark datasets TICaM [19] and TIMo [20] were only recently introduced, several research publications have already used them for evaluation and reported quantitative evaluation scores on them. In the following, we attempt to summarize and compare the heterogeneous results from all these publications as well as possible.…”