驾驶员动作检测综述
闵晨阳 2020.02.01
图片分类
多流融合的CNN
原文见参考文献【1】
datasets
SEU-DRIVING dataset，state-farm distracted driving dataset
Network architecture
Fusion strategy
Experiment results
state farm
SEU
多尺度注意机制CNN
原文见参考文献【2】
datasets
State-Farm,S-DA,R-DA
Network architecture
Experiment results
state farm
R-DA
S-DA
融合手脸特征的CNN
原文见参考文献【5】
datasets
AUC([5]首创)
Network architecture
Experiment results
视频分类
Two-Stream Inflated 3D ConvNet(I3D)
原文见参考文献【3】【4】
datasets
AUC,State Farm
经过对以上两个数据集的每一帧图片的组合得到20,094 10-frame clips for State
Farm and 14,536 10-frame clips for AUC dataset
Network architecture
Experiment results
基于空间-时间特征的CNN
原文见参考文献【6】
datasets
AUC ,Brain4Cars
Network architecture
数据融合：通过每一帧图片后增加几帧光流来联系“连续几帧图片”
融合后的四帧数据输入带有Softmax层(批量归一化层）的inception网络，每一帧结果融合后经过softmax层输出预测结果
Experiment results
多类型细粒度Drive&Act数据集
原文见参考文献【7】
datasets
Experiment results
Mid-Level
Action-Object-Location
Long-running Task
多类型数据融合的结果
参考文献
【1】Hu Y, Lu M, Lu X, et al. Driving behaviour recognition from still images by using multi-stream fusion CNN[J]. machine vision applications, 2019, 30(5): 851-865.
【2】Hu Y, Lu M, Lu X, et al. Feature refinement for image-based driver action recognition via multi-scale attention convolutional neural network[J]. Signal Processing-image Communication, 2020.
【3】Moslemi N, Azmi R, Soryani M, et al. Driver Distraction Recognition using 3D Convolutional Neural Networks[C]. international conference on pattern recognition, 2019.
【4】 J. Carreira and A. Zisserman, “Quo Vadis, Action Recognition? A New Model and the Kinetics Dataset,” in IEEE Conference on Computer Vision and Pattern Recognition (CVPR), 2017.
【5】Abouelnaga Y, Eraqi H M, Moustafa M, et al. Real-time Distracted Driver Posture Classification.[J]. arXiv: Computer Vision and Pattern Recognition, 2017.
【6】Kose N, Kopuklu O, Unnervik A, et al. Real-Time Driver State Monitoring Using a CNN Based Spatio-Temporal Approach*[C]. international conference on intelligent transportation systems, 2019.
【7】Martin M, Roitberg A, Haurilet M, et al. Drive&Act: A Multi-Modal Dataset for Fine-Grained Driver Behavior Recognition in Autonomous Vehicles[C]. international conference on computer vision, 2019: 2801-2810.