Conditional Random Fields as Recurrent Neural Networks

• Main Authors: Shuai Zheng, Jayasumana, Sadeep, Romera-Paredes, Bernardino, Vineet, Vibhav, Zhizhong Su, Dalong Du, Chang Huang, Torr, Philip H. S.
• Format: Conference Proceeding
• Language: English
• Subjects: Computer vision; Fields (mathematics); Graphical models; Image segmentation; Labeling; Labelling; Machine learning; Modelling; Networks; Neural networks; Segmentation; Semantics; Training
• ISSN: 2380-7504
• DOI: 10.1109/ICCV.2015.179

Pixel-level labelling tasks, such as semantic segmentation, play a central role in image understanding. Recent approaches have attempted to harness the capabilities of deep learning techniques for image recognition to tackle pixel-level labelling tasks. One central issue in this methodology is the limited capacity of deep learning techniques to delineate visual objects. To solve this problem, we introduce a new form of convolutional neural network that combines the strengths of Convolutional Neural Networks (CNNs) and Conditional Random Fields (CRFs)-based probabilistic graphical modelling. To this end, we formulate Conditional Random Fields with Gaussian pairwise potentials and mean-field approximate inference as Recurrent Neural Networks. This network, called CRF-RNN, is then plugged in as a part of a CNN to obtain a deep network that has desirable properties of both CNNs and CRFs. Importantly, our system fully integrates CRF modelling with CNNs, making it possible to train the whole deep network end-to-end with the usual back-propagation algorithm, avoiding offline post-processing methods for object delineation. We apply the proposed method to the problem of semantic image segmentation, obtaining top results on the challenging Pascal VOC 2012 segmentation benchmark.