Deep Spatio-Temporal Random Fields for Efficient Video Segmentation

2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition(2018)

引用 75|浏览96
暂无评分
摘要
In this work we introduce a time- and memory-efficient method for structured prediction that couples neuron decisions across both space at time. We show that we are able to perform exact and efficient inference on a densely connected spatio-temporal graph by capitalizing on recent advances on deep Gaussian Conditional Random Fields (GCRFs). Our method, called VideoGCRF is (a) efficient, (b) has a unique global minimum, and (c) can be trained end-to-end alongside contemporary deep networks for video understanding. We experiment with multiple connectivity patterns in the temporal domain, and present empirical improvements over strong baselines on the tasks of both semantic and instance segmentation of videos.
更多
查看译文
关键词
VideoGCRF,video segmentation,deep Gaussian conditional random fields,time-efficient method,neuron decisions,structured prediction,memory-efficient method,deep spatio-temporal random fields,video understanding,contemporary deep networks,densely-connected spatio-temporal graph
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