http://chineseinput.net/에서 pinyin(병음)방식으로 중국어를 변환할 수 있습니다.
변환된 중국어를 복사하여 사용하시면 됩니다.
CycleGAN 기반 데이터 증강을 이용한 딥러닝 기반 안개 감지 성능 개선 기법
류태광(Taekwang Ryu),이보원(Bowon Lee) 대한전자공학회 2020 대한전자공학회 학술대회 Vol.2020 No.11
In deep learning, the amount and quality of data have a significant impact on performance. A large amount of fog data and normal data are required for fog detection on deep learning-based methods. Normal data can be easily collected using black box when driving on normal situations, but fog conditions are not easy to acquire. In this paper, we generate fog data using CycleGAN and use this data to improve the performance of CNN-based fog detection models. It also shows that data augmentation using CycleGAN improves the generalization performance of the model.