http://chineseinput.net/에서 pinyin(병음)방식으로 중국어를 변환할 수 있습니다.
변환된 중국어를 복사하여 사용하시면 됩니다.
Deep Neural Network를 이용한 교통사고 과실비율 자동측정 시스템
최진우(Jinwoo Choi),이유노(Yunoh Lee),박인규(In Kyu Park) 대한전자공학회 2020 대한전자공학회 학술대회 Vol.2020 No.8
This paper proposes an automatic measurement system for negligence rate of car accident. The system takes the CCTV video of a car accident as input, then the negligence rate is automatically estimated using a series of deep neural networks. The system consists of 4 stages. First, we employ the YOLO and Deepsort models to find and track the visible cars in CCTV video. Second, it uses the Mask R—CNN model to find out which cars are involved to the accident. Third, the 3D CNN model classifies what kind of accident happens. Finally, using the random forest model, which gives the negligence percentage of the accident, the accident percentage of each car is estimated and displayed.