Efficient Driver Anomaly Detection via Conditional Temporal Proposal and Classification Network

Su, Lang; Cao, Dongpu; Sun, Chen

doi:10.1109/tcss.2022.3158480

Cited by 11 publications

(6 citation statements)

References 40 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Also, Hu et al [13] proposed a novel clustering supervised contrastive loss to optimize the distribution of the extracted representation vectors to improve the model performance. Su et al [33] proposed a two-stage anomaly detection and classification framework that ensures the model's understanding of NDRA deep features as well as robustness to Open-Set anomalies. However, the NDRA classification task's usually have a high background variation and a weak driver subject variation.…”

Section: Video Anomaly Detectionmentioning

confidence: 99%

“…In the experiments of the overall approach, we include Deep SVDD [52], Deep EPCC [53], DAD [12], DC-EPCC [54], CL-MOD [50], DADCNet [33], ConvGRU [55], DCHS [56], SuMoCo [57], and TDAD [58]were used as comparative algorithms, where DAD [12] was the baseline model, and comparison experiments were carried out on the 2 views and the 2 modalities in the DAD dataset as well as the comparison experiments were conducted on the fused views. The specific experimental results are shown in Tab.…”

Section: Comparison Experimentsmentioning

confidence: 99%

See 1 more Smart Citation

DSTANet: Learning a Dual-Stream Model for Anomaly Driving Action Detection Using Spatio-Temporal and Appearance Features

Peng,

Hu,

et al. 2024

Preprint

View full text Add to dashboard Cite

Driving action anomaly detection based on in-cab surveillance video has become the mainstream of current driving action research. However, there is a substantial redundancy of spatio-temporal information in the spatio-temporal action features extracted using only the 3D encoder, which will weaken the distinguishability between normal and anomaly videos, and there will be more difficult samples to be distinguished during the anomaly detection. To alleviate this problem, we propose a dual-stream model combining spatio-temporal and appearance features for anomaly driving action detection. Firstly, the model utilises 3D ResNet and 2D ResNet for feature extraction of spatio-temporal action and appearance information in the video. Next, to fully combine the respective advantages of the cross-dimension information and obtain fused features that can be more easily distinguished, a contrastive learning based alignment approach was used before fusion. In addition, cross-dimensional features are fused by introducing Transformer's cross attention block. Moreover, the supervised contrastive learning objective function is employed to guides the model to distinguish between normal and anomaly driving action. Finally, calculating the feature similarity between the target sample and the normal memory center is utilized to obtain the anomaly score of the target sample. In the experiment, the proposed model DSTANet in this paper achieved the 96.81% AUC in the publicly available DAD dataset, which is higher than the popular models in recent years. Furthermore, feature distinguishability was significantly improved in the visualization experiments. The experiment code and data are publicly available at: https://github.com/CreatedTRYNA/DCC\_Driving\_Anomaly\_Detection-.

show abstract

Section: Video Anomaly Detectionmentioning

confidence: 99%

Section: Comparison Experimentsmentioning

confidence: 99%

DSTANet: Learning a Dual-Stream Model for Anomaly Driving Action Detection Using Spatio-Temporal and Appearance Features

Peng,

Hu,

et al. 2024

Preprint

View full text Add to dashboard Cite

show abstract

“…Various multiview multimodal methods have been proposed with different emphases. Some propose novel learning methods (e.g., supervised contrastive learning [15]), while others [1,4,[21][22][23] are focused on handling the temporal dimension. However, how to combine heterogenous data in DMS has rarely been studied.…”

Section: Multimodal Driver Monitoring Systemsmentioning

confidence: 99%

“…However, how to combine heterogenous data in DMS has rarely been studied. Most previous methods [15,18,23] adopt a decision-level fusion by averaging the scores, while Ortega et al [21] propose to fuse data at an input level by concatenation. These strategies cannot handle modality/view interaction well and hence tend to underperform.…”

Section: Multimodal Driver Monitoring Systemsmentioning

confidence: 99%

Robust Multiview Multimodal Driver Monitoring System Using Masked Multi-Head Self-Attention

Ma,

Sanchez,

Nikan

et al. 2023

2023 IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops (CVPRW)

View full text Add to dashboard Cite

show abstract

“…Therefore, distractions can be identified from the driver's gaze or their behavior of secondary driving tasks. In particular, developing various existing sensors can help obtain the driver's gaze and movements, through both of which we can then identify driver distraction states [8,9].…”

Section: Introductionmentioning

confidence: 99%

Classification of Driver Distraction Risk Levels: Based on Driver’s Gaze and Secondary Driving Tasks

et al. 2022

View full text Add to dashboard Cite

Driver distraction is one of the significant causes of traffic accidents. To improve the accuracy of accident occurrence prediction under driver distraction and to provide graded warnings, it is necessary to classify the level of driver distraction. Based on naturalistic driving study data, distraction risk levels are classified using the driver’s gaze and secondary driving tasks in this paper. The classification results are then combined with road environment factors for accident occurrence prediction. Two ways are suggested to classify driver distraction risk levels in this study: one is to divide it into three levels based on the driver’s gaze and the AttenD algorithm, and the other is to divide it into six levels based on secondary driving tasks and odds ratio. Random Forest, AdaBoost, and XGBoost are used to predict accident occurrence by combining the classification results, driver characteristics, and road environment factors. The results show that the classification of distraction risk levels helps improve the model prediction accuracy. The classification based on the driver’s gaze is better than that based on secondary driving tasks. The classification method can be applied to accident risk prediction and further driving risk warning.

show abstract

Efficient Driver Anomaly Detection via Conditional Temporal Proposal and Classification Network

Cited by 11 publications

References 40 publications

DSTANet: Learning a Dual-Stream Model for Anomaly Driving Action Detection Using Spatio-Temporal and Appearance Features

DSTANet: Learning a Dual-Stream Model for Anomaly Driving Action Detection Using Spatio-Temporal and Appearance Features

Robust Multiview Multimodal Driver Monitoring System Using Masked Multi-Head Self-Attention

Classification of Driver Distraction Risk Levels: Based on Driver’s Gaze and Secondary Driving Tasks

Contact Info

Product

Resources

About