Learning shape priors for object segmentation via neural networks

Simon Safar, Ming-Hsuan Yang

Abstract

We present a joint algorithm for object segmentation that integrates both global shape and local edge information in a deep learning framework. The proposed architecture uses convolutional layers to extract image features, followed by a fully connected section to represent shapes specific to a given object class. This preliminary mask is further refined by matching segmentation mask patches to local features. These processing steps facilitate learning the shape priors effectively with a feedforward pass rather than complex inference methods. Furthermore, our novel convolutional refinement stage presents a convincing alternative to Conditional Random Fields, with promising results on multiple datasets.
BibTeX
@inproceedings{safar2015learning,
  author = {Simon Safar and Ming-Hsuan Yang},
  title = {{Learning shape priors for object segmentation via neural networks}},
  booktitle = {ICIP},
  year = {2015},
  doi = {10.1109/icip.2015.7351118},
}

Loading related content...

Figure