End-to-end semantic face segmentation with conditional random fields as convolutional, recurrent and adversarial networks
Computer Vision and Pattern Recognition
2017-03-10 v1
Abstract
Recent years have seen a sharp increase in the number of related yet distinct advances in semantic segmentation. Here, we tackle this problem by leveraging the respective strengths of these advances. That is, we formulate a conditional random field over a four-connected graph as end-to-end trainable convolutional and recurrent networks, and estimate them via an adversarial process. Importantly, our model learns not only unary potentials but also pairwise potentials, while aggregating multi-scale contexts and controlling higher-order inconsistencies. We evaluate our model on two standard benchmark datasets for semantic face segmentation, achieving state-of-the-art results on both of them.
Cite
@article{arxiv.1703.03305,
title = {End-to-end semantic face segmentation with conditional random fields as convolutional, recurrent and adversarial networks},
author = {Umut Güçlü and Yağmur Güçlütürk and Meysam Madadi and Sergio Escalera and Xavier Baró and Jordi González and Rob van Lier and Marcel A. J. van Gerven},
journal= {arXiv preprint arXiv:1703.03305},
year = {2017}
}