Abstract:Point cloud segmentation is the foundation of 3D environmental perception for modern intelligent systems. To solve this problem and image segmentation, conditional random fields (CRFs) are usually formulated as discrete models in label space to encourage label consistency, which is actually a kind of postprocessing. In this paper, we reconsider the CRF in feature space for point cloud segmentation because it can capture the structure of features well to improve the representation ability of features rather than simply smoothing. Therefore, we first model the point cloud features with a continuous quadratic energy model and formulate its solution process as a message-passing graph convolution, by which it can be easily integrated into a deep network. We theoretically demonstrate that the message passing in the graph convolution is equivalent to the mean-field approximation of a continuous CRF model. Furthermore, we build an encoder-decoder network based on the proposed continuous CRF graph convolution (CRFConv), in which the CRFConv embedded in the decoding layers can restore the details of high-level features that were lost in the encoding stage to enhance the location ability of the network, thereby benefiting segmentation. Analogous to the CRFConv, we show that the classical discrete CRF can also work collaboratively with the proposed network via another graph convolution to further improve the segmentation results. Experiments on various point cloud benchmarks demonstrate the effectiveness and robustness of the proposed method. Compared with the state-of-the-art methods, the proposed method can also achieve competitive segmentation performance.

Global Ranking of Documents Using Continuous Conditional Random Fields

Global Ranking Using Continuous Conditional Random Fields

CRF with Locality-Consistent Dictionary Learning for Semantic Segmentation

Automatic Indexing Model Based on Conditional Random Fields

Conditional Topic Random Fields.

Semi-supervised document retrieval

Continuous Conditional Random Field Convolution for Point Cloud Segmentation

Conditional Random Fields for Image Region Labeling with Global Observation.

Learning to rank based on user relevance feedback

An Attention-Based Interactive Learning-to-Rank Model for Document Retrieval

Efficient robust conditional random fields

Improving User's Preferences Ranking from Click-Through Data

Learning Groupwise Multivariate Scoring Functions Using Deep Neural Networks

Reinforcement Learning to Rank with Markov Decision Process

Semi-supervised training for conditional random fields with pseudo auxiliary task.

Feature selection method based on conditional relevance

Ranking with Query Influence Weighting for Document Retrieval

Information Extraction from Chinese Research Papers Based on Conditional Random Fields

Learning to Co-Embed Queries and Documents

Discriminative Training of Deep Fully-connected Continuous CRF with Task-specific Loss

A Noise-Tolerant Graphical Model for Ranking