Pistachio: Towards Synthetic, Balanced, and Long-Form Video Anomaly Benchmarks

TL;DR

Pistachio introduces a synthetic, controllable video benchmark for anomaly detection and understanding, addressing limitations of existing datasets by providing diverse, complex, and long-form videos generated with minimal manual effort.

Contribution

A novel generation-based pipeline for creating scalable, diverse, and complex video benchmarks for anomaly detection and understanding tasks.

Findings

Pistachio provides 41-second coherent videos with diverse anomalies.

The benchmark reveals new challenges for current anomaly detection methods.

Extensive experiments demonstrate Pistachio's scale, diversity, and complexity.

Abstract

Automatically detecting abnormal events in videos is crucial for modern autonomous systems, yet existing Video Anomaly Detection (VAD) benchmarks lack the scene diversity, balanced anomaly coverage, and temporal complexity needed to reliably assess real-world performance. Meanwhile, the community is increasingly moving toward Video Anomaly Understanding (VAU), which requires deeper semantic and causal reasoning but remains difficult to benchmark due to the heavy manual annotation effort it demands. In this paper, we introduce Pistachio, a new VAD/VAU benchmark constructed entirely through a controlled, generation-based pipeline. By leveraging recent advances in video generation models, Pistachio provides precise control over scenes, anomaly types, and temporal narratives, effectively eliminating the biases and limitations of Internet-collected datasets. Our pipeline integrates scene-conditioned anomaly assignment, multi-step storyline generation, and a temporally consistent long-form synthesis strategy that produces coherent 41-second videos with minimal human intervention. Extensive experiments demonstrate the scale, diversity, and complexity of Pistachio, revealing new challenges for existing methods and motivating future research on dynamic and multi-event anomaly understanding.