资源分类
》 资源 》论文》Deep Learning Shape Priors for Object Segmentation

Deep Learning Shape Priors for Object Segmentation

2019-11-28 | |  77 |   39 |   0

Abstract  In this paper we introduce a new shape-driven approach  for object segmentation. Given a training set of shapes, we  first use deep Boltzmann machine to learn the hierarchical  architecture of shape priors. This learned hierarchical  architecture is then used to model shape variations of  global and local structures in an energetic form. Finally, it  is applied to data-driven variational methods to perform  object extraction of corrupted data based on shape  probabilistic representation. Experiments demonstrate that  our model can be applied to dataset of arbitrary prior  shapes, and can cope with image noise and clutter, as well  as partial occlusions.

上一篇：Label-Embedding for Attribute-Based Classification

下一篇：Studying Relationships BetweenHuman Gaze, Description, and Computer Vision

用户评价
全部评价

热门资源

  • The Variational S...

    Unlike traditional images which do not offer in...

  • Learning to Predi...

    Much of model-based reinforcement learning invo...

  • Stratified Strate...

    In this paper we introduce Stratified Strategy ...

  • A Mathematical Mo...

    Direct democracy, where each voter casts one vo...

  • Rating-Boosted La...

    The performance of a recommendation system reli...