ETRI-Knowledge Sharing Plaform

ENGLISH

성과물

논문 검색
구분 SCI
연도 ~ 키워드

상세정보

학술대회 Driver Distraction Detection using Single Convolutional Neural Network
Cited 25 time in scopus Download 7 time Share share facebook twitter linkedin kakaostory
저자
김휘, 최현균, 장병태, 임진수
발행일
201710
출처
International Conference on Information and Communication Technology Convergence (ICTC) 2017, pp.1204-1206
DOI
https://dx.doi.org/10.1109/ICTC.2017.8190898
초록
Driver status detection is an essential task because driver distraction, fatigue, and drowsiness of driver are serious causes of traffic accident in recent. In this paper, we focus on driver distraction and propose a method to detect driver distraction. We detect driver distraction using single Convolutional Neural Network model such as Inception ResNet and MobileNet. As our experiments, both models can be trained with a small amount of dataset and checkpoints which were pre-trained with ILSVRC2012 dataset. Furthermore, although our training dataset consists images of two subjects, our method shows reliable result for other subjects.
키워드
Driver Distraction Detection, Driver Status Detection, Inception Resnet, MobileNet
KSP 제안 키워드
Convolution neural network(CNN), Distraction detection, Driver distraction, Driver status, Status detection, Traffic accident, neural network model