TY  - JOUR
PY  - 2022//
TI  - Improvement of vehicle position estimation using CNN-based vehicle bottom face center detection
JO  - Transactions of the Korean Society of Automotive Engineers
A1  - Kim, Gahyun
A1  - Jung, Ho Gi
A1  - Suhr, Jae Kyu
SP  - 599
EP  - 607
VL  - 30
IS  - 7
N2  - This paper proposes a method to improve vehicle position accuracy by detecting its bottom face center(BFC) based on a convolutional neural network. The proposed method is implemented by simply adding only the heads for BFC detection after maintaining the architecture of the existing vehicle detector. The BFC is calculated based on the feature map obtained from the vehicle detector. In order for the network to detect the BFC, the origin and distance function must be defined. In this paper, eight combinations of two methods for the origin, and four methods for the distance function were compared and evaluated. The performance of the proposed method is quantitatively evaluated by using Euclidean distance error and normalized Euclidean distance error. The result of this study revealed that the proposed method shows a 94.6 % improvement in vehicle position accuracy, compared to the previous method.     키워드: 차량 위치 추정, 밑면 중심점, 딥뉴럴네트워크, V2I 통신, 감시 카메라<p />  <p>Language: ko</p>
LA  - ko
SN  - 1225-6382
UR  - http://dx.doi.org/10.7467/KSAE.2022.30.7.599
ID  - ref1
ER  -