Street Scene Video Anomaly Detection Dataset

<p><strong><span>Introduction</span></strong></p> <p><span>The Street Scene dataset consists of 46 training video sequences and 35 testing video sequences taken from a static USB camera looking down on a scene of a two-lane street with bike lanes and pedestrian sidewalks.<span>&nbsp; </span>Videos were collected from the camera at various times during two consecutive summers.<span>&nbsp; </span>All of the videos were taken during the daytime.<span>&nbsp; </span>The dataset is challenging because of the variety of activities taking place such as cars driving, turning, stopping and parking; pedestrians walking, jogging and pushing strollers; and bikers riding in bike lanes. In addition, the videos contain changing shadows, and moving background such as a flag and trees blowing in the wind.</span></p> <p><span>There are a total of 202,545 color video frames (56,135 for training and 146,410 for testing) each of size 1280 x 720 pixels. The frames were extracted from the original videos at 15 frames per second.</span></p> <p><span>The 35 testing sequences have a total of 205 anomalous events consisting of 17 different anomaly types. A complete list of anomaly types and the number of each in the test set can be found in our paper.</span></p> <p><span>Ground truth annotations are provided for each testing video in the form of bounding boxes around each anomalous event in each frame. Each bounding box is also labeled with a track number, meaning each anomalous event is labeled as a track of bounding boxes. Track lengths vary from tens of frames to 5200 which is the length of the longest testing sequence. A single frame can have more than one anomaly labeled.</span></p> <p><span>NOTE: This version of the dataset differs slightly with the original made available in 2020.<span>&nbsp; </span>Some anomalies were found in a few of the normal training sequences.<span>&nbsp; </span>These training frames were deleted from the dataset.<span>&nbsp; </span>Specifically, the following frames were removed:</span></p> <p><span>Train026: frames 1-184 (car taking a u-turn)</span></p> <p><span>Train027: frames 1-229 (jay walkers)</span></p> <p><span>Train031: frames 1-299 (jay walkers, illegally parked car)</span></p> <p><strong><span>At a Glance</span></strong></p> <ul> <li><span>The size of the unzipped dataset is ~46GB</span></li> <li><span>The dataset consists of Train sequences (containing only videos with normal activity), Test sequences (containing some anomalous activity) along with ground truth annotations, and a README.md file describing the data organization and ground truth annotation format.</span></li> <li><span>The zip file contains a Train directory, a Test directory and a README.md file.</span></li> </ul> <p><strong><span>Other Resources</span></strong></p> <p><span>None</span></p> <p><strong><span>Citation</span></strong></p> <p><span>If you use the Street Scene dataset in your research, please cite our contribution:</span></p> <pre><code>@inproceedings{ramachandra2020street, title={Street Scene: A new dataset and evaluation protocol for video anomaly detection}, author={Ramachandra, Bharathkumar and Jones, Michael}, booktitle={Proceedings of the IEEE/CVF Winter Conference on Applications of Computer Vision}, pages={2569--2578}, year={2020} } </code></pre> <p><strong><span>License</span></strong></p> <p><span>The Street Scene dataset is released under&nbsp;</span><a href="https://creativecommons.org/licenses/by-sa/4.0/"><span>CC-BY-SA-4.0 license</span></a><span>.</span></p> <p><span>All data:</span></p> <pre><code>Created by Mitsubishi Electric Research Laboratories (MERL), 2023 SPDX-License-Identifier: CC-BY-SA-4.0 </code></pre>