The Research on Image Segmentation Based on the Minimum Error Probability Bayesian Decision Theory

The image segmentation technology has been extensively applied in many fields. As the foundation of image identification, the effective image segmentation plays a significant role during the course of subsequent image processing. Many theories and methods have been presented and discussed about image segmentation, such as K-means and fuzzy C-means methods, method based on regions information, method based on image edge detection, etc. In this work, it is proposed to apply Bayesian decision-making theory based on minimum error probability to gray image segmentation. The approach to image segmentation can guarantee the segmentation error probability minimum, which is generally what we desire. On the assumption that the gray values accord with the probability distribution of Gaussian finite mixture model in image feature space, EM algorithm is used to estimate the parameters of mixture model. In order to improve the convergence speed of EM algorithm, a novel method called weighted equal interval sampling is presented to obtain the contracted sample set. Consequently, the computation burden of EM algorithm is greatly reduced. The final experiments demonstrate the feasibility and high effectiveness of the method.

Download Full-text

The Research on Image Segmentation for Fingerprint Identification

Advanced Materials Research ◽

10.4028/www.scientific.net/amr.217-218.396 ◽

2011 ◽

Vol 217-218 ◽

pp. 396-401

Author(s):

Xiao Jie Xu ◽

Xi Yan Dong

Keyword(s):

Image Segmentation ◽

Mixture Model ◽

Hybrid Approach ◽

Identification Accuracy ◽

Fingerprint Identification ◽

Image Feature ◽

Background Information ◽

Bayesian Decision ◽

Fingerprint Image ◽

Directional Information

As the precondition of fingerprint identification, the effective image segmentation plays the significant role in the following image processing. Unlike other images, the fingerprint images are obviously directional. Aiming at this feature, in this paper, an image segmentation method based on the directional information of fingerprint image is introduced, which sufficiently utilizes the directional information of fingerprint image and succeeds in separating the background information. However, owing to the absence of directional information in some local areas of fingerprint image, this method will produce large segmentation errors, even fail. Therefore, for these local regions without directional information, it is proposed to apply Bayesian decision-making theory based on minimum error probability to realize image segmentation. On the assumption that the gray values accord with the probability distribution of Gaussian finite mixture model in image feature space, EM algorithm is used to estimate the parameters of mixture model. The mixture application of two methods can effectively separate the background information from fingerprint image while saving the preprocessing time and ensuring the following identification accuracy of fingerprint. The experiments illustrate the feasibility of the hybrid approach.

Download Full-text

Recognizing Cursive Typewritten Text Using Segmentation-Free System

The Scientific World JOURNAL ◽

10.1155/2015/818432 ◽

2015 ◽

Vol 2015 ◽

pp. 1-7 ◽

Cited By ~ 2

Author(s):

Mohammad S. Khorsheed

Keyword(s):

Feature Extraction ◽

Feature Vector ◽

Feature Space ◽

Recognition System ◽

Image Feature ◽

Text Recognition ◽

Statistical Features ◽

Free System ◽

Novel Method ◽

Image Pixels

Feature extraction plays an important role in text recognition as it aims to capture essential characteristics of the text image. Feature extraction algorithms widely range between robust and hard to extract features and noise sensitive and easy to extract features. Among those feature types are statistical features which are derived from the statistical distribution of the image pixels. This paper presents a novel method for feature extraction where simple statistical features are extracted from a one-pixel wide window that slides across the text line. The feature set is clustered in the feature space using vector quantization. The feature vector sequence is then injected to a classification engine for training and recognition purposes. The recognition system is applied to a data corpus which includes cursive Arabic text of more than 600 A4-size sheets typewritten in multiple computer-generated fonts. The system performance is compared to a previously published system from the literature with a similar engine but a different feature set.

Download Full-text

Color Image Segmentation Using Gaussian Mixture Model and EM Algorithm

Multimedia and Signal Processing - Communications in Computer and Information Science ◽

10.1007/978-3-642-35286-7_9 ◽

2012 ◽

pp. 61-66 ◽

Cited By ~ 11

Author(s):

Zhaoxia Fu ◽

Liming Wang

Keyword(s):

Image Segmentation ◽

Em Algorithm ◽

Gaussian Mixture Model ◽

Mixture Model ◽

Color Image ◽

Gaussian Mixture ◽

Color Image Segmentation

Download Full-text

Nonrigid point set registration based on Laplace mixture model with local constraints

Assembly Automation ◽

10.1108/aa-06-2019-0108 ◽

2019 ◽

Vol 40 (2) ◽

pp. 335-343

Author(s):

Chao Xu ◽

Xianqiang Yang ◽

Xiaofeng Liu

Keyword(s):

Em Algorithm ◽

Mixture Model ◽

Gaussian Mixture ◽

Nonrigid Registration ◽

Model Parameters ◽

The Novel ◽

Content Type ◽

Point Set Registration ◽

Point Set ◽

Novel Method

Purpose This paper aims to investigate a probabilistic mixture model for the nonrigid point set registration problem in the computer vision tasks. The equations to estimate the mixture model parameters and the constraint items are derived simultaneously in the proposed strategy. Design/methodology/approach The problem of point set registration is expressed as Laplace mixture model (LMM) instead of Gaussian mixture model. Three constraint items, namely, distance, the transformation and the correspondence, are introduced to improve the accuracy. The expectation-maximization (EM) algorithm is used to optimize the objection function and the transformation matrix and correspondence matrix are given concurrently. Findings Although amounts of the researchers study the nonrigid registration problem, the LMM is not considered for most of them. The nonrigid registration problem is considered in the LMM with the constraint items in this paper. Three experiments are performed to verify the effectiveness and robustness and demonstrate the validity. Originality/value The novel method to solve the nonrigid point set registration problem in the presence of the constraint items with EM algorithm is put forward in this work.

Download Full-text

DCS-SVM: a novel semi-automated method for human brain MR image segmentation

Biomedical Engineering / Biomedizinische Technik ◽

10.1515/bmt-2015-0226 ◽

2017 ◽

Vol 62 (6) ◽

pp. 581-590 ◽

Cited By ~ 2

Author(s):

Ali Ahmadvand ◽

Mohammad Reza Daliri ◽

Mohammadtaghi Hajiali

Keyword(s):

Image Segmentation ◽

Computational Cost ◽

Three Dimensional ◽

Feature Space ◽

Support Vector ◽

Brain Images ◽

Automated Method ◽

Classifier Selection ◽

Novel Method ◽

Mr Brain

AbstractIn this paper, a novel method is proposed which appropriately segments magnetic resonance (MR) brain images into three main tissues. This paper proposes an extension of our previous work in which we suggested a combination of multiple classifiers (CMC)-based methods named dynamic classifier selection-dynamic local training local Tanimoto index (DCS-DLTLTI) for MR brain image segmentation into three main cerebral tissues. This idea is used here and a novel method is developed that tries to use more complex and accurate classifiers like support vector machine (SVM) in the ensemble. This work is challenging because the CMC-based methods are time consuming, especially on huge datasets like three-dimensional (3D) brain MR images. Moreover, SVM is a powerful method that is used for modeling datasets with complex feature space, but it also has huge computational cost for big datasets, especially those with strong interclass variability problems and with more than two classes such as 3D brain images; therefore, we cannot use SVM in DCS-DLTLTI. Therefore, we propose a novel approach named “DCS-SVM” to use SVM in DCS-DLTLTI to improve the accuracy of segmentation results. The proposed method is applied on well-known datasets of the Internet Brain Segmentation Repository (IBSR) and promising results are obtained.

Download Full-text

Novel Method of MRI Medical Image Segmentation Combining Watershed Algorithm and WKFCM Algorithm

Applied Mechanics and Materials ◽

10.4028/www.scientific.net/amm.121-126.4518 ◽

2011 ◽

Vol 121-126 ◽

pp. 4518-4522

Author(s):

Jin Qing Liu ◽

Kun Chen

Keyword(s):

Image Segmentation ◽

Clustering Algorithm ◽

Brain Mri ◽

Feature Space ◽

Medical Image Segmentation ◽

Watershed Algorithm ◽

Kernel Clustering ◽

Mercer Kernel ◽

Novel Method ◽

Segmentation Problem

Aimed at the disadvantage of over-segmentation that traditional watershed algorithm segmented MRI images, a new method of MRI image segmentation was presented. First, through traditional watershed segmentation algorithm, the image was segmented into different areas, and then based on the improved kernel-clustering algorithm, we used Mercer-kernel to map average gray value of each area to high-dimensional feature space, making originally not displayed features manifested. In this way, we can achieve a more accurate clustering, and solve over-segmentation problem of watershed algorithm segmenting MRI images efficiently, thereby get better segmentation result. Experimental results show that the method of this paper can segment brain MRI images satisfactorily, and obtain clearer segmentation images.

Download Full-text