谷歌浏览器插件
订阅小程序
在清言上使用

ActionFlowNet: Learning Motion Representation for Action Recognition.

2018 IEEE WINTER CONFERENCE ON APPLICATIONS OF COMPUTER VISION (WACV 2018)(2018)

引用 86|浏览129
关键词
motion information,data-efficient representation learning approach,raw pixels,convolutional neural networks,unlabeled videos,unsupervised representation learning methods,external large scale data,external labeled datasets,video representation learning,action recognition,single stream network training,ImageNet datasets,Sport-1M datasets,motion representation learning,multitask learning model,ActionFlowNet,optical flow input
AI 理解论文
溯源树
样例
生成溯源树,研究论文发展脉络
Chat Paper
正在生成论文摘要