Extended Markov Random Fields for Predictive Image Segmentation

Abstract

Since the 1970s, there has been increasing interest in the use of Markov Random Fields (MRFs) as models to aid in the segmentation of noisy or degraded digital images. MRFs can make up for deficiencies in observed information by adding a-priori knowledge to the image interpretation process in the form of models of spatial interaction between neighbouring pixels. In data fusion problems, interaction might also be assumed between corresponding pixels in two different kinds of image of the same scene. Alternatively, temporal interaction might be assumed between corresponding pixels in consecutive frames of a video sequence. In object tracking or robotic navigation problems, a similar relationship may exist between pixels of an observed image and those of a predicted image, derived from models of the motion and scene. In all of these cases the MRF model can be extended to incorporate this additional knowledge. This paper explains the theory of Extended-Markov Random Field (E-MRF) segmentation techniques, surveys the research which has been crucial to their development and presents results from new work in this area with an application to robotic vision in conditions of extremely poor visibility.

Single image super-resolution reconstruction based on genetic algorithm and regularization prior model

“…edge priors [43], gradient priors [42], steering kernel regression (SKR) [56,28], non-local means (NLMs) [4,29,22] and total variation (TV) [30]. Alternative ideas about prior knowledge and reconstruction constraints use Markov Random Fields (MRF) to impose probabilistic constraints on pixel consistency [3,17], and have been extended to combine these consistency constraints with predicted or expected image content [38,39,40]. Prior knowledge methods have proved effective at suppressing noise and preserving edges.…”

Bayesian fusion of thermal and visible spectra camera data for region based tracking with rapid background adaptation

“…In [5], a human pedestrian is initially segmented from high resolution visible spectrum images using background subtraction. The segmented image is then modified by combining it with opinions from pixels of a low resolution thermal image, expressing the fusion problem probabilistically as an Extended-Markov Random Field, [6]. [1] formulates a statistical background model for each pixel, similar to the well known surveillance-tracking formulation of [7], but including thermal parameters and colour parameters in a single, high dimensional distribution for each pixel.…”

Continuous Machine Learning in Computer Vision - Tracking with Adaptive Class Models

“…an ellipse or a square), a flexible contour or snake or the projection of a known 3D model of a tracked rigid body. An example of an algorithm which continuously relearns both object and background models, is the EM/E-MRF algorithm (Stolkin et al, 2000(Stolkin et al, , 2007b(Stolkin et al, , 2007c, which combines simple, Gaussian models of the object and background pixel intensities with a 3D CAD model of the rigid tracked object. The EM/E-MRF algorithm was an attempt to tackle images under conditions of extremely poor visibility, by combining observed image data with prior knowledge in various forms.…”