http://chineseinput.net/에서 pinyin(병음)방식으로 중국어를 변환할 수 있습니다.
변환된 중국어를 복사하여 사용하시면 됩니다.
다양한 입력 영상과 훈련 방법을 통한 인간-객체-상호작용의 성능 분석
이건우(Geon U Lee),조정찬(Jung Chan Cho) 대한전자공학회 2020 대한전자공학회 학술대회 Vol.2020 No.8
Understanding the relationship between a human and an object is very useful in real life. However, current research on human-object-interaction (HOI) has not achieved practical performance. In this paper, we proposed a three-stream-based classification model and analyzed the effect of the input combination of human, object, interaction, and union images. Our experimental results show that modeling the relationship between a human and an object is more important than which feature extraction network applies to images.