Towards Computationally Efficient and Realtime Distracted Driver Detection With MobileVGG Network

Baheti, Bhakti; Talbar, Sanjay N.; Gajre, Suhas

doi:10.1109/tiv.2020.2995555

Cited by 66 publications

(31 citation statements)

References 34 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The results showed that the proposed model had similar performance to other state-of-the-art methods. The study in [11] introduced a computationally efficient distracted driver detection system based on convolutional neural networks. The authors proposed a new architecture called mo-bileVGG.…”

Section: Single-based Deep Learning Modelsmentioning

confidence: 99%

E2DR: A Deep Learning Ensemble-Based Driver Distraction Detection with Recommendations Model

Aljasim

Kashef

2022

Sensors

View full text Add to dashboard Cite

The increasing number of car accidents is a significant issue in current transportation systems. According to the World Health Organization (WHO), road accidents are the eighth highest top cause of death around the world. More than 80% of road accidents are caused by distracted driving, such as using a mobile phone, talking to passengers, and smoking. A lot of efforts have been made to tackle the problem of driver distraction; however, no optimal solution is provided. A practical approach to solving this problem is implementing quantitative measures for driver activities and designing a classification system that detects distracting actions. In this paper, we have implemented a portfolio of various ensemble deep learning models that have been proven to efficiently classify driver distracted actions and provide an in-car recommendation to minimize the level of distractions and increase in-car awareness for improved safety. This paper proposes E2DR, a new scalable model that uses stacking ensemble methods to combine two or more deep learning models to improve accuracy, enhance generalization, and reduce overfitting, with real-time recommendations. The highest performing E2DR variant, which included the ResNet50 and VGG16 models, achieved a test accuracy of 92% as applied to state-of-the-art datasets, including the State Farm Distracted Drivers dataset, using novel data splitting strategies.

show abstract

Section: Single-based Deep Learning Modelsmentioning

confidence: 99%

E2DR: A Deep Learning Ensemble-Based Driver Distraction Detection with Recommendations Model

Aljasim

Kashef

2022

Sensors

View full text Add to dashboard Cite

show abstract

“…We selected several approaches [23,[35][36][37][38][39][40] from the literature proposed for detecting distracted driving behaviors for the comparison with our proposed HSDDD framework. J. M. Mase et al [23] have presented a driver distraction posture detection method in which first CNNs are leveraged for automatically learning the spatial posture features and then stacked Bidirectional Long Short-Term Memory (BiLSTM) Networks are used to capture the spectral-spatio features of the images by extracting the spectral features amongst the stacked feature maps from the pre-trained CNNs.…”

Section: Comparison With Existing Workmentioning

confidence: 99%

“…Moslemi et al [37] have presented a detection technique for the distracted drivers which is based on 3D CNN, and optical flow is utilized so that the detection accuracy can be improved. B. Baheti et al [38] have proposed a CNNbased approach and have developed a new architecture, named mobileVGG, based on depth-wise separable convolutions and VGG16. A. Ezzouhri et al [39] have used deeplearning-based segmentation for extracting the driver's body parts, before performing the distraction detection and classification task.…”

Section: Comparison With Existing Workmentioning

confidence: 99%

HSDDD: A Hybrid Scheme for the Detection of Distracted Driving through Fusion of Deep Learning and Handcrafted Features

Alkinani

Khan

Arshad

et al. 2022

Sensors

View full text Add to dashboard Cite

Traditional methods for behavior detection of distracted drivers are not capable of capturing driver behavior features related to complex temporal features. With the goal to improve transportation safety and to reduce fatal accidents on roads, this research article presents a Hybrid Scheme for the Detection of Distracted Driving called HSDDD. This scheme is based on a strategy of aggregating handcrafted and deep CNN features. HSDDD is based on three-tiered architecture. The three tiers are named as Coordination tier, Concatenation tier and Classification tier. We first obtain HOG features by using handcrafted algorithms, and then at the coordination tier, we leverage four deep CNN models including AlexNet, Inception V3, Resnet50 and VGG-16 for extracting DCNN features. DCNN extracted features are fused with HOG extracted features at the Concatenation tier. Then PCA is used as a feature selection technique. PCA takes both the extracted features and removes the redundant and irrelevant information, and it improves the classification performance. After feature fusion and feature selection, the two classifiers, KNN and SVM, at the Classification tier take the selected features and classify the ten classes of distracted driving behaviors. We evaluate our proposed scheme and observe its performance by using the accuracy metrics.

show abstract

“…In this subsection, we also compare the confusion matrix and the class-wise sensitivity between the baseline and the proposed method. As defined in [51], [52], class-wise sensitivity ( ), which is also known as the true positive rate (TPR), is computed as…”

Section: A Comparison With the Baselinesmentioning

confidence: 99%

TML: A Triple-Wise Multi-Task Learning Framework for Distracted Driver Recognition

et al. 2021

View full text Add to dashboard Cite

We propose a multi-task learning framework for improving the performance of vision-based deep-learning approaches for driver distraction recognition. The most popular tool so far for solving this task is convolutional neural networks (CNNs) that have proven to be strongly biased toward local features. Such bias causes CNNs to neglect global structural information, adversely affecting the robustness of the distracted driver recognition task. To solve this problem, we generate positive and negative samples of each given input, and construct a triplet of images (i.e., raw image, positive sample, and negative sample). The positive sample is generated by applying structure-aware illumination to the human body region of each given input. The negative sample is generated by randomly shuffling the local regions of each given input. The networks are then trained with the triplets using a multi-task learning strategy to force the networks to explore global information by multiple tasks: (a) recognizing the raw input and positive sample as the given ground truth; (b) recognizing the negative sample as an extra "meaningless" label; (c) pulling closer the distance between the features obtained from the raw input and positive sample while pushing away the distance between the features obtained from the raw input and negative sample. By doing so, the model can be trained so that it neglects the background information and pays more attention to the global structual information of the scene. The proposed approach reaches state-of-the-art performance on the AUC Distracted Driver Dataset and performs better than state-of-the-art studies on the Drive and Act Dataset. With raw images as input, we have achieved an accuracy of 96.0% for the AUC distracted driver dataset and 66.8% for the Drive and Act Dataset. Our approach does not introduce extra overhead during the testing procedure (i.e., utilization procedure), which is helpful for real-life applications. Moreover, better accuracy can be achieved by fusing the predictions respectively obtained with the raw input and positive sample. As a result, we have achieved an accuracy of 96.3% for the AUC distracted driver Dataset and 66.9% for the Drive and Act Dataset. The class activation map (CAM) of our proposed method is subjectively more reasonable, which would enhance the reliability and explainability of the model. INDEX TERMSAction recognition, Advanced driver assistance, Contrastive learning, Multi-task learning, Intelligent vehicles I. INTRODUCTIONNowadays, distracted driving has become a huge threat to human society. According to the report issued by the National Highway Traffic Safety Administration (NHTSA) in the United State in 2019, traffic accidents caused by distracted driving led to 3,142 or 8.7 percent of all accidents of this year in the United States [1], and most of them were involved in texting or talking on mobile phones. Owing to this situation, a reduction in traffic accidents can be realized if we can develop distracted driving detectors. Such detectors can be use...

show abstract

Towards Computationally Efficient and Realtime Distracted Driver Detection With MobileVGG Network

Cited by 66 publications

References 34 publications

E2DR: A Deep Learning Ensemble-Based Driver Distraction Detection with Recommendations Model

E2DR: A Deep Learning Ensemble-Based Driver Distraction Detection with Recommendations Model

HSDDD: A Hybrid Scheme for the Detection of Distracted Driving through Fusion of Deep Learning and Handcrafted Features

TML: A Triple-Wise Multi-Task Learning Framework for Distracted Driver Recognition

Contact Info

Product

Resources

About