Recognizing abnormal events such as traffic violations and accidents in natural driving scenes is essential for successful autonomous driving and advanced driver assistance systems. However, most work on video anomaly detection suffers from two crucial drawbacks. First, they assume cameras are fixed and videos have static backgrounds, which is reasonable for surveillance applications but not for vehicle-mounted cameras. Second, they pose the problem as one-class classification, relying on arduously hand-labeled training datasets that limit recognition to anomaly categories that have been explicitly trained. This paper proposes an unsupervised approach for traffic accident detection in first-person (dashboard-mounted camera) videos. Our major novelty is to detect anomalies by predicting the future locations of traffic participants and then monitoring the prediction accuracy and consistency metrics with three different strategies. We evaluate our approach using a new dataset of diverse traffic accidents, AnAn Accident Detection (A3D), as well as another publicly-available dataset. Experimental results show that our approach outperforms the state-of-the-art. Code and the dataset developed in this work are

Predicting the future location of vehicles is essential for safety-critical applications such as advanced driver assistance systems (ADAS) and autonomous driving. This paper introduces a novel approach to simultaneously predict both the location and scale of target vehicles in the first-person (egocentric) view of an ego-vehicle. We present a multi-stream recurrent neural network (RNN) encoder-decoder model that separately captures both object location and scale and pixellevel observations for future vehicle localization. We show that incorporating dense optical flow improves prediction results significantly since it captures information about motion as well as appearance change. We also find that explicitly modeling future motion of the ego-vehicle improves the prediction accuracy, which could be especially beneficial in intelligent and automated vehicles that have motion planning capability. To evaluate the performance of our approach, we present a new dataset of first-person videos collected from a variety of scenarios at road intersections, which are particularly challenging moments for prediction because vehicle trajectories are diverse and dynamic. Code and dataset have been made available at: https: //usa.honda-ri.com/hevi arXiv:1809.07408v2 [cs.CV]

No abstract

Autonomous vehicles require reliable and resilient sensor suites and ongoing validation through fleet-wide data collection. This paper proposes a Smart Black Box (SBB) to augment traditional low-bandwidth data logging with valuedriven high-bandwidth data capture. The SBB caches shortterm histories of data as buffers through a deterministic Mealy machine based on data value and similarity. Compression quality for each frame is determined by optimizing the trade-off between value and storage cost. With finite storage, prioritized data recording discards low-value buffers to make room for new data. This paper formulates SBB compression decision making as a constrained multi-objective optimization problem with novel value metrics and filtering. The SBB has been evaluated on a traffic simulator which generates trajectories containing events of interest (EOIs) and corresponding first-person view videos. SBB compression efficiency is assessed by comparing storage requirements with different compression quality levels and event capture ratios. Performance is evaluated by comparing results with a traditional first-in-first-out (FIFO) recording scheme. Deep learning performance on images recorded at different compression levels is evaluated to illustrate the reproducibility of SBB recorded data.

No abstract