Towards Anomaly Detection in Dashcam Videos
This addresses the problem of improving driver safety systems for dashcam users by moving beyond simple model-driven algorithms, though it is incremental as it adapts existing anomaly detection ideas to a new domain with limited prior work.
The paper tackles the problem of detecting anomalies in dashcam videos, where existing hand-crafted methods are inadequate due to the long-tailed distribution of road hazards, by proposing data-driven deep learning approaches and introducing a new dataset called RetroTrucks, with experiments showing the dataset is more challenging and previous methods perform poorly.
Inexpensive sensing and computation, as well as insurance innovations, have made smart dashboard cameras ubiquitous. Increasingly, simple model-driven computer vision algorithms focused on lane departures or safe following distances are finding their way into these devices. Unfortunately, the long-tailed distribution of road hazards means that these hand-crafted pipelines are inadequate for driver safety systems. We propose to apply data-driven anomaly detection ideas from deep learning to dashcam videos, which hold the promise of bridging this gap. Unfortunately, there exists almost no literature applying anomaly understanding to moving cameras, and correspondingly there is also a lack of relevant datasets. To counter this issue, we present a large and diverse dataset of truck dashcam videos, namely RetroTrucks, that includes normal and anomalous driving scenes. We apply: (i) one-class classification loss and (ii) reconstruction-based loss, for anomaly detection on RetroTrucks as well as on existing static-camera datasets. We introduce formulations for modeling object interactions in this context as priors. Our experiments indicate that our dataset is indeed more challenging than standard anomaly detection datasets, and previous anomaly detection methods do not perform well here out-of-the-box. In addition, we share insights into the behavior of these two important families of anomaly detection approaches on dashcam data.