视频异常检测——Learning Memory-guided Normality for Anomaly Detection

文章

2020.CVPR.Learning Memory-guided Normality for Anomaly Detection.pdf

资源

github
homepage

问题描述

基于reconstruction的问题：
1、 do not consider the diversity of normal patterns explicitly,
2、powerful representation capacity of CNNs allows to reconstruct abnormal video frames

思想

1、use a memory module record prototypical patterns of normal data to lessening the capacity of CNNs.
2、feature compactness loss:mapping the features of a normal video frame to the nearest item in the memory and encouraging them to be close
separateness loss:minimizes the distance between each feature and its nearest item, while maximizing the discrepancy between the feature and the second nearest one

贡献

We propose to use multiple prototypes to represent the diverse patterns of normal video frames for unsupervised anomaly detection. To this end, we introduce a memory module recording prototypical patterns of normal data on the items in the memory.
We propose feature compactness and separateness losses to train the memory, ensuring the diversity and discriminative power of the memory items. We also present a new update scheme of the memory, when both normal and abnormal samples exist at test time.
We achieve a new state of the art on standard benchmarks for unsupervised anomaly detection in video sequences. We also provide an extensive experimental analysis with ablation studies.

做法

loss：

异常分数：

存在的问题

M是固定的，可能真的prototype和实际的数量不符。

思考

encoder部分可能等于提取出某些特征（比如柱子，人，台阶），然后将这些东西分成M类，测试的时候，某一个离prototype远，就出问题。