Revisiting Deep Structured Models for Pixel-Level Labeling with Gradient-Based Inference
Journal article, 2018

Semantic segmentation and other pixel-level labeling tasks have made significant progress recently due to the deep learning paradigm. Many state-of-the-art structured prediction methods also include a random field model with a hand-crafted Gaussian potential to model spatial priors and label consistencies and feature-based image conditioning. These random field models with image conditioning typically require computationally demanding filtering techniques during inference. In this paper, we present a new inference and learning framework which can learn arbitrary pairwise conditional random field (CRF) potentials. Both standard spatial and high-dimensional bilateral kernels are considered. In addition, we introduce a new type of potential function which is image-dependent like the bilateral kernel, but an order of magnitude faster to compute since only spatial convolutions are employed. It is empirically demonstrated that such learned potentials can improve segmentation accuracy and that certain label-class interactions are indeed better modeled by a non-Gaussian potential. Our framework is evaluated on several public benchmarks for semantic segmentation with improved performance compared to previous state-of-the-art CNN+CRF models.

deep learning

semantic segmentation

deep structured models

conditional random fields


Måns Larsson

Chalmers, Electrical Engineering, Signal Processing and Biomedical Engineering

Anurag Arnab

University of Oxford

Shuai Zheng

University of Oxford

Philip Torr

University of Oxford

Fredrik Kahl

Chalmers, Electrical Engineering, Signal Processing and Biomedical Engineering

SIAM Journal on Imaging Sciences

19364954 (eISSN)

Vol. 11 4 2610-2628

Subject Categories

Other Computer and Information Science

Probability Theory and Statistics

Computer Vision and Robotics (Autonomous Systems)



More information

Latest update