Skip to content

Dataset · Liyun1009

Video anomaly detection benchmark (NeurIPS 2024)

Liyun1009/MSAD

A collection of real-world RGB videos for evaluating anomaly detection algorithms across various locations.

What's inside

A video dataset containing fewer than 1000 clips, presented at the NeurIPS 2024 conference in the Datasets and Benchmarks track. It features real color RGB video recordings from various indoor rooms and street scenes. The material is intended for studying and evaluating methods for detecting atypical events in video.

What you can do with it

  • Testing video anomaly detection models for surveillance systems
  • Comparing algorithms for finding abnormal situations in indoor and outdoor environments
Size
up to 1K rows
Downloads this month
737
Likes
11
Access
on request on Hugging Face

How to load it

from datasets import load_dataset

ds = load_dataset("Liyun1009/MSAD")