a2c [1910.03695v1] NADS-Net: A Nimble Architecture for Driver and Seat Belt Detection via Convolutional Neural Networks

Enable JavaScript to see more content

arXiv2code // top new 14d 1m 2m 3m // Enable JavaScript to see more content

[share]	Go to arXiv [ U of Iowa ]Download as Jupyter Notebook	2019-10-12
[1910.03695v1] NADS-Net: A Nimble Architecture for Driver and Seat Belt Detection via Convolutional Neural Networks NADS-Net is capable of estimating body pose together with seat belt segmentation with the similar accuracy than the state of the art baseline [4], while requiring fewer parameters and demonstrating a faster inference time Abstract A new convolutional neural network (CNN) architecture for 2D driver/passenger pose estimation and seat belt detection is proposed in this paper. The new architecture is more nimble and thus more suitable for in-vehicle monitoring tasks compared to other generic pose estimation algorithms. The new architecture, named NADS-Net, utilizes the feature pyramid network (FPN) backbone with multiple detection heads to achieve the optimal performance for driver/passenger state detection tasks. The new architecture is validated on a new data set containing video clips of 100 drivers in 50 driving sessions that are collected for this study. The detection performance is analyzed under different demographic, appearance, and illumination conditions. The results presented in this paper may provide meaningful insights for the autonomous driving research community and automotive industry for future algorithm development and data collection. ‹Figure 1: First and third columns in the figure are results produced by NADS-Net and second and fourth columns are the corresponding ground truth annotations. (Related works)Figure 2: Data collection setup for this study. (Data set)Figure 3: NADS-Net architecture. (Data set)Figure 4: Modified PCKh metric (mPCKh) used for key point evaluation. (Result and discussion)Figure 5: Seat belt detection result (left) in comparison with the ground truth (right). With human inspection, the prediction result is of good quality as it correctly marks the seat belt area. However, IOU for this particular example was 46%, justifying the need for a better evaluation metric. (Result and discussion)Figure 6: Saliency map visualization [31] for the ‘right wrist’ class. Note visual cues from the face have significant contribution to the prediction. (Appearances)›