视频中的异常检测是一个具有挑战性的计算机视觉问题。现有的最先进视频异常检测方法主要集中在深度神经网络的结构设计上，以获得性能改进。与主要研究趋势不同，本文侧重于将集成学习和深度神经网络相结合，提出了一种基于集成生成对抗网络（Generative Adversarial Networks，GAN）的方法。在所提出的方法中，一组生成器和一组判别器一起训练，因此每个生成器可以从多个判别器获得反馈，反之亦然。与单个GAN相比，集成GAN可以更好地对正常数据的分布进行建模，从而更好地检测异常。在两个公开数据集上测试了所提出的方法性能。结果表明，集成学习显著提高了单个检测模型的性能，在两个数据集上比现有最近方法分别超过0.4%和0.3%的帧级AUC。

Anomaly detection in video is one of the challenging computer vision problems. The existing state-of-the-art video anomaly detection methods mainly focus on the structural design of deep neural networks to obtain performance improvements. Different from the main research trend, this article focuses on the combination of ensemble learning and deep neural network, and proposes a method based on ensemble generative adversarial networks (GAN). In the proposed method, a set of generators and discriminators are trained together, so each generator gets feedback from multiple discriminators, and vice versa. Compared with a single GAN, an ensemble GAN can better model the distribution of normal data, thereby better detecting anomalies. The performance of the proposed method is tested on two public data sets. The results show that ensemble learning significantly improves the performance of a single detection model, and the performance of ensemble GAN exceeds the frame-level AUC of 0.4% and 0.3% on the two data sets compared with the existing recent methods, respectively.