# Street Scene: A new dataset and evaluation protocol for video anomaly detection

## TL;DR

This paper introduces a large, diverse dataset called Street Scene and new evaluation protocols to improve video anomaly detection research, along with two improved baseline algorithms demonstrating superior performance.

## Contribution

The paper provides a new large-scale dataset and evaluation criteria, along with enhanced baseline algorithms, addressing current limitations in video anomaly detection research.

## Key findings

- The new dataset is larger and more varied than existing ones.
- Proposed evaluation criteria better estimate real-world performance.
- Baseline algorithms outperform existing state-of-the-art methods on Street Scene.

## Abstract

Progress in video anomaly detection research is currently slowed by small datasets that lack a wide variety of activities as well as flawed evaluation criteria. This paper aims to help move this research effort forward by introducing a large and varied new dataset called Street Scene, as well as two new evaluation criteria that provide a better estimate of how an algorithm will perform in practice. In addition to the new dataset and evaluation criteria, we present two variations of a novel baseline video anomaly detection algorithm and show they are much more accurate on Street Scene than two state-of-the-art algorithms from the literature.