Surveillance video anomaly detection is to detect events that rarely or never happened in a certain scene. The generation error (GE)-based methods exhibit excellent performances on this task. They firstly train a generative neural network (GNN) to generate normal samples, then judge the samples with large GEs as anomalies. Almost all the GE-based methods utilize frame-level GEs to detect anomalies. However, anomalies generally occur in local areas, the frame-level GE introduces GEs of normal areas to anomaly detection, that brings two problems: i) The GEs of normal areas reduce the anomaly saliency of the anomalous frame. ii) Different videos have different normal-GE-levels, thus it is hard to set a uniform threshold for different videos to detect anomalies. To address these problems, we propose a promotion method: utilize the maximum of block-level GEs on the frame to detect anomalies. Firstly, we calculate the block-level GEs at each position on the frame. Then, we utilize the maximum of the block-level GEs on the frame to detect anomalies. Based on the existing GNN models, we carry out experiments on multiple datasets. The results demonstrate the effectiveness of the proposed method and achieve state-of-the-art performance.

监视视频异常检测用于检测在特定场景中很少或从未发生的事件。基于生成错误（GE）的方法在此任务上表现出出色的性能。他们首先训练一个生成神经网络（GNN）以生成正常样本，然后将具有大GEs的样本判断为异常。几乎所有基于GE的方法都利用帧级GE来检测异常。然而，异常通常发生在局部区域，框架级GE将正常区域的GE引入异常检测，这带来了两个问题：i）正常区域的GE降低了异常框架的异常显着性。ii）不同的视频具有不同的正常GE级别，因此很难为不同的视频设置统一的阈值以检测异常。为了解决这些问题，我们提出了一种升级方法：利用帧上最大的块级GE来检测异常。首先，我们计算帧上每个位置的块级GE。然后，我们利用帧中最大的块级GE来检测异常。基于现有的GNN模型，我们对多个数据集进行了实验。结果证明了所提出方法的有效性并获得了最新的性能。