MAVAD: Audio-Visual Dataset and Method for Anomaly Detection in Traffic Videos
Authors:
- 27 October 2024
- conference paper
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 1106-1112
Abstract
This paper introduces the first audio-visual dataset for traffic anomaly detection called MAVAD, taken from real-world scenes, with a diverse range of illumination conditions. In addition, a novel anomaly detection method is proposed which combines visual and audio features extracted from video sequences by means of cross-attention. We demonstrate that the addition of audio improves anomaly detection performance by up to 5.2%. Moreover, the impact of image anonymization is evaluated, showing only a minor decrease in performance averaging at 1.7%.Keywords
All Related Versions
This publication has 21 references indexed in Scilit:
- Single-layer vision transformers for more accurate early exits with less overheadNeural Networks, 2022
- Spatial Data and Intelligence2022
- Learning to detect anomaly events in crowd scenes from synthetic dataNeurocomputing, 2021
- Multi-Modal Anomaly Detection by Using Audio and Visual CuesIEEE Access, 2021
- Not only Look, But Also Listen: Learning Multimodal Violence Detection Under Weak Supervision2020
- Graph Convolutional Label Noise Cleaner: Train a Plug-And-Play Action Classifier for Anomaly Detection2019
- Future Frame Prediction for Anomaly Detection - A New Baseline2018
- Real-World Anomaly Detection in Surveillance Videos2018
- Quo Vadis, Action Recognition? A New Model and the Kinetics Dataset2017
- Microsoft COCO: Common Objects in Context2014