The recognition of incorrect fastening of seat belts is significant in passenger and driver safety for the automotive industry and public health. It should be made sure that the passenger’s seat belt is not only fastened but also correctly fastened across the body so that the passenger is adequately protected in the event of an accident. Current technology employs the buckle effect sensor, which merely solves the buckling detection problem, but there is no reliable solution for the correct positioning of the seat belt. Additionally, computer vision-based systems are still incapable of recognizing the incorrect positioning of seat belts when the training is performed by employing the subjects out of the fleet. Considering this fact, in this study, we propose a novel solution that employs a vision-based incorrect fastening seat belt detector to perform automatic vertical height adjustment independent from drivers and passengers for the fleet vehicles. We recognize the incorrect positioning of the seat belt inside the car by the acceptable distance of the seat belt from the neck of drivers or passengers to avoid neck injuries and the deaths caused by neck cuts. An extensive benchmarking is performed by comparing the three CNN architectures such as; DenseNet121, GoogLeNet (Inception-v3), ResNet50 with respect to sensitivity, specificity, precision, false-positive rate, false-negative rate, F1 score, and accuracy. Additionally, training and validation loss curves and accuracy curves are plotted for all the models. Later, the three models are evaluated with a precision-recall (PR) curve at the end. According to the results, the DenseNet121 achieved the highest classification accuracy among the tested models with 99.95%. This paper includes information about the proposed system elements, registration of data, elaboration of data, program algorithm, testing the system in the lab, and on the vehicle.
AkayDKurtM. Otomobil Emniyet Kemeri Kullanılabilirlik Testi [Automobile Seat Belt Usability Test]. Gazi Universitesi Muhendislik Mimarlık Fakultesi Dergisi [Gazi Univ J Eng Architect Faculty]2014; 21(1): 183–191.
3.
KohliGAathiMKSethiM. Road accidents among adolescents. J Asia Pac Stud2013; 3(2): 245–259.
KaplanBOzcebeH. Trafik Kazaları Ve Arka Koltuk Guvenligi [Traffic accidents and back seat safety]. Toplum Hekimligi Bulteni [Bull Commun Med]2009; 28(1): 1–7.
10.
NordfjærnTJørgensenSHRundmoT. Safety attitudes, behaviour, anxiety and perceived control among professional and non-professional drivers. J Risk Res2012; 15(8): 875–896.
11.
DeliceM. Investigation of the effect of use of safety belt rate on the number of death and injury in traffic accidents. J Hum Sci2015; 12(1): 392–416.
12.
EvansL. Restraint effectiveness, occupant ejection from cars, and fatality reductions. Accid Anal Prevent1990; 22(2): 167–175.
13.
PickrellTMLiuC. Seat belt use in 2013 overall results. Traffic Safety Facts Research Note. Report No. DOT HS 811 875, 2014. National Highway Traffic Safety Administration.
Hodson-WalkerNJ. The value of safety belts: a review. Can Med Assoc J1970; 102: 391–393.
17.
CummingsP. Association of seat belt use with death: a comparison of estimates based on data from police and estimates based on data from trained crash investigators. Inj Prev2002; 8: 338–341.
18.
DonaldsonWFHanksSENassrA, et al. Cervical spine injuries associated with the incorrect use of airbags in motor vehicle collisions. Spine (Phila Pa 1976)2008; 33: 631–634.
19.
DawsonLKJenkinsNH. Fatal intra-abdominal injury associated with incorrect use of a seat belt. J Accid Emerg Med1998; 15: 437–438.
20.
GreenbaumEHarrisLHalloranWX. Flexion fracture of the lumbar spine due to lap-type seat belts. Calif Med1970, 113:74-6.
21.
SabeyBGrantBHobbsC. Alleviation of injuries by use of seat belts. Melbourne: Department of the Environment Department of Transport. https://trl.co.uk/sites/default/files/SR289.pdf (accessed 27 January 2020).
22.
DongZWuYPeiM, et al. Vehicle type classification using a semi-supervised convolutional neural network. IEEE Trans Intell Transp Syst2015; 16(4): 2247–2256.
23.
HuangYWuRSunY, et al. Vehicle logo recognition system based on convolutional neural networks with a pretraining strategy. IEEE Trans Intell Transp Syst2015; 16(4): 1951–1960.
24.
FangJZhouYYuY, et al. Fine-grained vehicle model recognition using a coarse-to-fine convolutional neural network architecture. IEEE Trans Intell Transp Syst2016; 18(7): 1782–1792.
25.
GuoHLinHZhangS, et al. Image-based seat belt detection. In: Proceedings of ICVES, Beijing, China, 10–12 July 2011, pp. 161–164. New York: IEEE.
26.
ArtanYBulanOLoceRP, et al. Passenger compartment violation detection in hov/hot lanes. IEEE Trans Intell Transp Syst2016; 17(2): 395–405.
27.
BaltaxeMMerguiRNistelK, et al. Marker-less vision-based detection of improper seat belt routing. In: 2019 IEEE intelligent vehicles symposium (IV), pp.783–789. New York: IEEE.
28.
HuangGLiuZVan der MateenL, et al. Densely connected convolutional networks. In: Proceedings of IEEE conference on computer vision and pattern recognition, July 2017, pp.2261–2269. New York: IEEE.
29.
SzegedyCLiuWJiaY, et al. Going deeper with convolutions. In: Proceedings of IEEE conference on computer vision and pattern recognition, June2015, pp.1–9. New York: IEEE.
30.
HeKZhangXRenS, et al. Deep residual learning for image recognition. In: Proceedings of IEEE conference on computer vision and pattern recognition, July 2016, pp.770–778. New York: IEEE.
BloiceMRothMPHolzingerA. Biomedical image augmentation using Augmenter. Bioinformatics2019; 35(21): 4522–4524.
33.
SimonyanKSizzermanA. Very deep convolutional networks for large-scale image recognition. arXiv preprint arXiv:1409.1556, 2014.
34.
BochkovskiyAWangCYLiaoHYM. YOLOv4: optimal speed and accuracy of object detection. arXiv preprint arXiv:2004.10934, 2020.
35.
HeKGkioxariGDollarP, et al. “Mask R-CNN.” In: Proceedings of IEEE conference on computer vision and pattern recognition, October 2017, pp.2961–2969. New York: IEEE.
36.
JustusDBrennanJBonnerS, et al. Predicting the computational cost of deep learning models. In: IEEE international conference on big data, December 2018, pp.3873–3882. New York: IEEE.
37.
LeCunYBottouLBengioY, et al. Gradient-based learning applied to document recognition. Proc IEEE1998; 86(11): 2278–2324. New York: IEEE.
38.
RumelhartDEHintonGEWilliamsRJ. Learning representations by back-propagating errors. Nature1986; 323(6088): 533–536.
39.
GoodfellowIPouget-AbadieJMirzaM, et al. Generative adversarial nets. arXiv preprint arXiv:1406.2661, 2014.
40.
GoodfellowIBengioYCourvilleA, et al. Deep learning, Vol. 1, no. 2. Cambridge, MA: MIT Press, 2016.
41.
SergeyISzegedC. Batch normalization: accelerating deep network training by reducing internal covariate shift. In: Proceedings of the 32nd international conference on machine learning – Volume 37 (ICML’15). JMLR.org, Lille, France, 2015, pp.448–456. Brookline, MA: Microtome Publishing.
42.
GlorotXBordesABengioY. Deep sparse rectifier neural networks. In: Proceedings of the 14th international conference on artificial intelligence and statistics (AISTATS) 2011, Vol. 15, Fort Lauderdale, FL, USA, 14 June 2011, pp.315–323. Brookline, MA: Microtome Publishing.
43.
GauswamiMTrivediK. Implementation of machine learning for gender detection using CNN on raspberry Pi platform. In: 2018 2nd International conference on inventive systems and control (ICISC). Coimbatore, India, 2018. New York: IEEE.
