Statistical Modeling of Deep Features to Reduce False Alarms in Video Change Detection

Detecting relevant changes is a fundamental problem of video surveillance. Because of the high variability of data and the difficulty of properly annotating changes, unsupervised methods dominate the field. Arguably one of the most critical issues to make them practical is to reduce their false alarm rate. In this work, we develop a non-semantic, method-agnostic, weakly supervised a-contrario validation process, based on high-dimensional statistical modeling of deep features using a Gaussian mixture model, that can reduce the number of false alarms of any change detection algorithm. We also raise the insufficiency of the conventionally used pixel-wise evaluation, as it fails to precisely capture the performance needs of most real applications. For this reason, we complement pixel-wise metrics with component-wise metrics and evaluate the impact of our approach at both pixel and object levels, on six methods and several sequences from different datasets. Our experimental results reveal that the a-contrario theory can be applied to a statistical model of the background of a scene and largely reduce the number of false positives at both pixel and component levels.