Image Segmentation Using Markov Random Field Model in Fully Parallel Cellular Network Architectures

doi:10.1006/rtim.1998.0159

Real-Time Imaging

Volume 6, Issue 3, June 2000, Pages 195-211

https://doi.org/10.1006/rtim.1998.0159 Get rights and content

Abstract

Markovian approaches to early vision processes need a huge amount of computing power. These algorithms can usually be implemented on parallel computing structures. Herein, we show that the Markovian labeling approach can be implemented in fully parallel cellular network architectures, using simple functions and data representations. This makes possible to implement our model in parallel imaging VLSI chips.

As an example, we have developed a simplified statistical image segmentation algorithm for the Cellular Neural/Nonlinear Networks Universal Machine (CNN-UM), which is a new image processing tool, containing thousands of cells with analog dynamics, local memories and processing units. The Modified Metropolis Dynamics (MMD) optimization method can be implemented into the raw analog architecture of the CNN-UM. We can introduce the whole pseudo-stochastic segmentation process in the CNN architecture using 8 memories/cell. We use simple arithmetic functions (addition, multiplication), equality-test between neighboring pixels and very simple nonlinear output functions (step, jigsaw). With this architecture, the proposed VLSI CNN chip can execute a pseudo-stochastic relaxation algorithm of about 100 iterations in about 100 μs.

In the suggested solution the segmentation is unsupervised, where a pixel-level statistical estimation model is used. We have tested different monogrid and multigrid architectures.

In our CNN-UM model several complex preprocessing steps can be involved, such as texture-classification or anisotropic diffusion. With these preprocessing steps, our fully parallel cellular system may work as a high-level image segmentation machine, using only simple functions based on the close-neighborhood of a pixel.

References (34)

A. Caplier et al.
Real-time implementations of an MRF-based motion detection algorithm
Real-Time Imaging
(1998)
F Heitz et al.
Multiscale minimization of global energy functions in some visual recovery problems
CVGIP: Image Understanding
(1994)
Z. Kato et al.
A Hierarchical Markov Random Field Model and Multitemperature Annealing for Parallel Classification
Graphical Models and Image Processing
(1996)
T. Szirányi et al.
Texture Classification and Segmentation by Cellular Neural Network using Genetic Learning
Computer Vision and Image Understanding
(1998)
Azencott, R. 1987, Markov fields and image analysis, Proc. AFCET, Antibes,...
R. Azencott
Parallel Simulated Annealing: Parallelization Techniques
(1992)
D.A. Bader et al.
Scalable data parallel algorithms for texture synthesis using Gibbs random fields
IEEE Tr. Image Processing
(1995)
J. Besag
On the statistical analysis of dirty pictures
Journal of Royal Statist. Soc.
(1996)
A. Blake
Comparison of the efficiency of deterministic and stochastic algorithms for visual reconstruction
IEEE Trans. on Pattern Analysis and Machine Intelligence
(1989)
C.A. Bouman et al.
A Multiscale Random Field Model for Bayesian Image Segmentation
IEEE Trans. Image Proc.
(1994)

F. Catte et al.

Image selective smoothing and edge detection by nonlinear diffusion

SIAM J. Numerical Anal.

(1992)

L.O. Chua et al.

Cellular Neural Networks

IEEE Trans. Circuit and Systems

(1998)

Czúni, L. Szirányi, T. 1996, Picture segmentation with an anisotropic preliminary step to an MRF model with Cellular...

L. Czúni et al.

Multigrid MRF Based Picture Segmentation with Cellular Neural Networks

Proc. CAIP’97, Lect. Notes in Comp. Sci.

(1997)

R. Dominguez-Castro et al.

A 0.8 mm CMOS 2-D Programmable Mixed-Signal Focal-Plane Array Processor with On-Chip Binary Imaging and Instructions Storage

IEEE J. Solid State Circuits

(1997)

S. Geman et al.

Stochastic relaxation, Gibbs distributions and the Bayesian restoration of images

IEEE Trans. Pattern Analysis and Machine Intelligence

(1984)

Cited by (38)

Segmentation and Feature Extraction in Medical Imaging: A Systematic Review
2020, Procedia Computer Science
Image processing techniques being crucial towards analyzing and resolving issues in medical imaging since last two decades. Medical imaging is a process or technique to find the inner or outer construction of mortal body. The process observes medicinal diagnosis, analyze illnesses and develop data-sets of normal and abnormal imageries. Medical imaging is divided in two folds such as invisible-light medical imaging and visible-light medical imaging. The second type of medical imaging were can be understood by a common person whereas the first type can be interpreted by a radiologist. Analysis of all these require segmentation and feature extraction. In fact a lot of medical imaging techniques are available but authors restrict survey to tumor detection through mammograms or magnetic resonance imaging. In this paper, authors survey on various segmentation and feature extraction methods in medicinal images used for preprocessing.
Probabilistic terrain classification in unstructured environments
2013, Robotics and Autonomous Systems
Citation Excerpt :
MRFs have been frequently used in this field, especially for segmentation tasks. Szirányi et al. [5] address the problem of the huge amount of computing power required for Markovian approaches.
Autonomous navigation in unstructured environments is a complex task and an active area of research in mobile robotics. Unlike urban areas with lanes, road signs, and maps, the environment around our robot is unknown and unstructured. Such an environment requires careful examination as it is random, continuous, and the number of perceptions and possible actions are infinite.
We describe a terrain classification approach for our autonomous robot based on Markov Random Fields (MRFs ) on fused 3D laser and camera image data. Our primary data structure is a 2D grid whose cells carry information extracted from sensor readings. All cells within the grid are classified and their surface is analyzed in regard to negotiability for wheeled robots.
Knowledge of our robot’s egomotion allows fusion of previous classification results with current sensor data in order to fill data gaps and regions outside the visibility of the sensors. We estimate egomotion by integrating information of an IMU, GPS measurements, and wheel odometry in an extended Kalman filter.
In our experiments we achieve a recall ratio of about 90% for detecting streets and obstacles. We show that our approach is fast enough to be used on autonomous mobile robots in real time.
A new fuzzy segmentation approach based on S-FCM type 2 using LBP-GCO features
2012, Signal Processing: Image Communication
Citation Excerpt :
In Fig. 18, the first row presents the original images. The second one illustrates the segmentation results introduced in Refs. [13,44], from where it can be noticed that there is some misclassification at different regions. Moreover, borders are not effectively identified.
Gabor filtering is a widely applied approach for texture analysis. This technique shows a strong dependence on certain number of parameters. Unfortunately, each variation of values of any parameter may affect the texture characterization performance. Moreover, Gabor filters are unable to extract micro-texture features which also have a negative effect on the clustering task. This paper, deals with a new descriptor which avoids the drawbacks mentioned above. The novel texture descriptor combines grating cell operator outputs derived from a designed Gabor filters bank, and local binary pattern features. For the clustering task, an extended version of fuzzy type 2 clustering algorithm is also proposed. The effectiveness of the proposed segmentation approach on a variety of synthetic and textured images is highlighted. Several experimental results on a set of textured images show the superiority of the proposed approach in terms of segmentation accuracy with respect to quantitative and qualitative comparisons.
Detection of soldering defects in Printed Circuit Boards with Hierarchical Marked Point Processes
2011, Pattern Recognition Letters
Citation Excerpt :
We performed our tests on a standard desktop computer, where processing 1 MP input images with 76 SP entities took around 10–20 s. Apart from further source code optimization, there are several options to accelerate the method from algorithmic points of view. Firstly, similar to other graph-like model structures (such as Markov Random Fields (Szirányi et al., 2000)), MPP optimization algorithms can be parallelized and adopted for multiprocessor architectures (Bambos and Chen, 1994). Secondly, in case of input images with sufficiently high Signal to Noise Ratio (SNR), the convergence of the MBD relaxation process can be significantly sped up with using so called “birth maps” as shown in (Descombes et al., 2009).
In this paper we introduce a probabilistic approach for optical quality checking of solder pastes (SP) in Printed Circuit Boards (PCB). Dealing with unregistered image inputs, our task is to address at the same time SP identification, and detection of special soldering errors, called scooping. For this reason we introduce a novel Hierarchical Marked Point Process (HMPP) framework, which is able to handle the paste and scooping extraction problems simultaneously, so that the SPs and included scoops have a parent–child relationship. A global optimization process attempts to find the optimal configuration of entities, considering the observed data, prior knowledge, and interactions between the neighboring circuit elements. The proposed method is evaluated on a real PCB image set containing more than 3000 SPs and 600 scooping artifacts. A morphology-based baseline method is also introduced for the problem and used as reference for qualitative and quantitative comparison against the proposed model.
Parallelized segmentation of a serially sectioned whole human brain
2008, Image and Vision Computing
Building a digital three dimensional representation of a human brain is a challenging task. Such a model provides insights into the microstructure of cortical layering and columns. The presented work is based on a complete dissected and preserved human brain that has been serially sectioned at a coronal resolution that is suitable for single cell detection. More than 6000 sections have been generated and exist as digital images. To obtain a valuable three dimensional representation, morphology preserving affine linear and nonlinear registration schemes are necessary steps. To rebuild a serially sectioned brain, reference images derived from a non deformed object, e.g., MRI or block face images, are necessary for a faithful affine linear and nonlinear registration. In the case of block face images the brain regions must be separated from highly variable background regions to obtain a suitable stack of segmentation images. Among the image segmentation algorithms we found fuzzy c-means techniques as a promising starting point for a sophisticated segmentation framework of either gray level or color images within 2- and 3-dimensions.
With respect to algorithmic complexity and computation cost, two fuzzy c-means algorithms were implemented. A proper image preprocessing strategy turned out to be necessary for accurate and robust segmentation results. Primarily, the algorithms work in a parametric resp. supervised mode. Additionally, an automatic mode helps to explore the parameter space within a reasonable range and to compare the segmentation result with an optimal one, provided by an expert. By minimizing the differences we can set up parameters that are used for series of adjacent images. So, it is possible to obtain optimal segmentations independent of illumination disturbances, artifacts and defocusing.
We present a complete high resolution and accurate segmentation of the first complete human brain that was sectioned, photographed and digitized at histologic resolution. Based on these images, a succeeding 3D representation is presented. Finally, a segmented and spatially correct straightened data set is available now for coregistration tasks together with the high resolution histologic data set.
Fast multiparametric electromagnetic full-wave inversion via solving contracting scattering data equations optimized by the 3-d mrf model
2020, IEEE Transactions on Microwave Theory and Techniques

View all citing articles on Scopus

^f1: [email protected]

^f2: [email protected]

^f3: [email protected]

^f4: [email protected]

^f5: [email protected]

View full text

Regular ArticleImage Segmentation Using Markov Random Field Model in Fully Parallel Cellular Network Architectures

Abstract

Real-Time Imaging

CVGIP: Image Understanding

Graphical Models and Image Processing

Computer Vision and Image Understanding

Parallel Simulated Annealing: Parallelization Techniques

Scalable data parallel algorithms for texture synthesis using Gibbs random fields

IEEE Tr. Image Processing

On the statistical analysis of dirty pictures

Journal of Royal Statist. Soc.

Comparison of the efficiency of deterministic and stochastic algorithms for visual reconstruction

IEEE Trans. on Pattern Analysis and Machine Intelligence

A Multiscale Random Field Model for Bayesian Image Segmentation

IEEE Trans. Image Proc.