Improved Gaussian Mixture Model Research Articles

Automatic Visual Detection System of Railway Surface Defects With Curvature Filter and Improved Gaussian Mixture Model

Rails are among the most important components of railway transportation, and real-time defects detection of the railway is an important and challenging task because of intensity inhomogeneity, low contrast, and noise. This paper presents an automatic railway visual detection system (RVDS) for surface defects and focuses on several key issues of RVDS. First, in view of challenges such as complex condition and orbital reflectance inequality, we put forward a region-of-interest detection region extraction algorithm by vertical projection and gray contrast algorithm. In addition, a curvature filter equipped with implicit computing and surface preserving power is studied to eliminate noise and keep only the details. Then, an improved fast and robust Gaussian mixture model based on Markov random field is established for accurate and rapid surface defect segmentation. Additionally, an expectation–maximization algorithm is applied to optimize the parameters. The experimental results demonstrate that the proposed method performs well with both noisy and railway images, which enables identification and segmentation of the defects from rail surface, achieving detection performance with 92% precision and 88.8% recall rate on average, and is robust compared with the related well-established approaches.

Robust video tracking algorithm: a multi‐feature fusion approach

This study proposes a novel robust video tracking algorithm consists of target detection, multi-feature fusion, and extended Camshift. Firstly, a novel target detection method that integrates Canny edge operator, three-frame difference, and improved Gaussian mixture model (IGMM)-based background modelling is provided to detect targets. The IGMM-based background modelling divides video frames into meshes to avoid pixel-wise processing. In addition, the output of the target detection is utilised to initialise the IGMM and to accelerate the convergence of iterations. Secondly, low-dimensional regional covariance matrices are introduced to describe video targets by fusing multiple features like pixel location, colour index, rotation and scale invariant features as well as uniform local binary patterns, and directional derivatives. Thirdly, an extended Camshift based on adaptive kernel bandwidth and robust H∞ state estimation is proposed to predict the states of fast moving targets and to reduce the mean shift iterations. Finally, the effectiveness of the proposed tracking algorithm is demonstrated via experiments.

Classification and Detection of Adulteration in Olive Oil Using Improved Gaussian Mixture Model and Regression by Artificial Bee Colony Algorithm

Gaussian mixture model (GMM) and Gaussian mixture regression (GMR) can be used to detect adulteration in extra virgin olive oil. The estimate of the GMM parameters is commonly obtained from the expectation- maximization (EM) algorithm. EM algorithm has some limitations such as local optimum problems and sensitivity to the initial values. In this paper, artificial bee colony (ABC) algorithm is used to determine the optimal parameters in GMM and GMR. To improve the optimized performance and reduce computational effort of ABC algorithm, the information sharing mechanism among the global best food sources is introduced in ABC. The improved GMM and GMR by artificial bee colony algorithm (GMMRABC) were used to discriminate and quantify the adulteration of extra virgin olive oil with rapeseed oil using FT-IR spectroscopy. It has been demonstrated that the proposed method is an accurate, rapid, stable strategy for identifying and quantifying the extra virgin olive oil.

An Improved Gaussian Mixture Model Method for Moving Object Detection

Aiming at the shortcomings of Gaussian mixture model background method, a moving object detection method mixed with adaptive iterative block and interval frame difference method in the Gaussian mixture model is proposed. In this method, the video sequences are divided into different size pieces in order to reduce the amount of calculation of the algorithm. It not only effectively solves the problem that the traditional Gaussian mixture model algorithm cannot detect large and slow moving object accurately, but also solves empty and no connection problems due to the introduction of block thought. The experimental results show that the improved algorithm has faster processing speed, better effect and better environment adaptability compared with the background of the Gaussian mixture model method. And it can detect moving object more accurately and completely.

An Improved Gaussian Mixture Model for Damage Propagation Monitoring of an Aircraft Wing Spar under Changing Structural Boundary Conditions.

Structural Health Monitoring (SHM) technology is considered to be a key technology to reduce the maintenance cost and meanwhile ensure the operational safety of aircraft structures. It has gradually developed from theoretic and fundamental research to real-world engineering applications in recent decades. The problem of reliable damage monitoring under time-varying conditions is a main issue for the aerospace engineering applications of SHM technology. Among the existing SHM methods, Guided Wave (GW) and piezoelectric sensor-based SHM technique is a promising method due to its high damage sensitivity and long monitoring range. Nevertheless the reliability problem should be addressed. Several methods including environmental parameter compensation, baseline signal dependency reduction and data normalization, have been well studied but limitations remain. This paper proposes a damage propagation monitoring method based on an improved Gaussian Mixture Model (GMM). It can be used on-line without any structural mechanical model and a priori knowledge of damage and time-varying conditions. With this method, a baseline GMM is constructed first based on the GW features obtained under time-varying conditions when the structure under monitoring is in the healthy state. When a new GW feature is obtained during the on-line damage monitoring process, the GMM can be updated by an adaptive migration mechanism including dynamic learning and Gaussian components split-merge. The mixture probability distribution structure of the GMM and the number of Gaussian components can be optimized adaptively. Then an on-line GMM can be obtained. Finally, a best match based Kullback-Leibler (KL) divergence is studied to measure the migration degree between the baseline GMM and the on-line GMM to reveal the weak cumulative changes of the damage propagation mixed in the time-varying influence. A wing spar of an aircraft is used to validate the proposed method. The results indicate that the crack propagation under changing structural boundary conditions can be monitored reliably. The method is not limited by the properties of the structure, and thus it is feasible to be applied to composite structure.

The embedded real-time detection system of moving object based on improved Gaussian mixture model

Because of its complexity, the Gaussian mixture model algorithm costs many computing resources (CPU cycles) and it is difficult to realise real-time moving object detection on the embedded system. A real-time motion detection method based on an improved Gaussian mixture model is presented in this paper, which is optimised and adjusted for the Gaussian mixture model. It is also optimised on C language and CPU for running on an embedded system. The design of software and hardware on DM6446 is presented in detail including the implementation of the algorithm. The experimental results show that the system can effectively overcome the interference from external environment and realise multi-object real-time motion detection and tracking.

Segmentation of Magnetic Resonance Brain Images Based on Improved Gaussian Mixture Model with Spatial Information

Segmentation of Magnetic Resonance Brain Images Based on Improved Gaussian Mixture Model with Spatial Information

Algorithm of Moving Vehicle Detection by Improved Gaussian Mixture Model Based on ROI

Detection of moving object is a hot topic in computer vision. Traditionally, it is detected for every pixel in whole image by Gaussian mixture background model, which may waste more time and space. In order to improving the computational efficiency, an advanced Gaussian mixture model based on Region of Interest was proposed. Firstly, the solution finds out the most probably region where the target may turn up. And then Gaussian mixture background model is built in this area. Finally, morphological filter algorithm is used for improving integrity of the detected targets. Results show that the improved method could have a more perfect detection but no more time increasing than typical method.

An Improved Gaussian Mixture Model based on NonLocal Information for Brain MR Images Segmentation

Brain image segmentation is an important part of medical image analysis. Due to the effect of imaging mechanism, MR images usually intensity in homogeneity, which is also named as bias field. Traditional Gaussian Mixed Model (GMM) method is hard to obtain satisfied segmentation results with the effect of noise and bias field. We propose a novel model based on GMM and nonlocal information. The improved method coupled segmentation and bias field correction that can manage the bias field while segmenting the image. In order to obtain a smooth bias field, we employed the Legendre Polynomials to fit it and merged it to the EM framework. We also use the non local information to deal with the noise and preserve geometrical edges information. The results show that our method can obtain more accurate results and bias field.

Characterization of Peanut Oil by Infrared Spectroscopy with an Improved Gaussian Mixture Model

The Gaussian mixture model (GMM) and regression (GMR) are widely used statistical tools in pattern classification and nonlinear regression. In this paper, the suitability score was used for variable selection to improve GMM and GMR. The improved GMM was used to characterize peanut oil adulterated with palm oil using Fourier transform infrared spectroscopy. The improved GMR was applied to determine the concentration of palm oil contaminant present. As comparison, GMM and GMR with principal component analysis for feature extraction, support vector machine, back-propagation artificial neural network, nearest centroid classification, and partial least-squares analysis were also used to classify and quantify peanut oil. It was demonstrated that the method is a new classification and regression strategy for the detection of adulterated edible oil.

An Improved Moving Body Detection Method Based on Gaussian Mixture Model and Differencing

This paper presented a detecting algorithm based on improved Gaussian mixture model, which improved the speed of establishing and updating the model. Whilst, according to the detection requirements in different time periods, using the method of combining temporal differencing and background differencing to detect moving body improved the instantaneity and veracity of moving body detection.

Fast and robust head detection with arbitrary pose and occlusion

Head detection in images and videos plays an important role in a wide range of computer vision and surveillance applications. Aiming to detect heads with arbitrarily occluded faces and head pose, in this paper, we propose a novel Gaussian energy function based algorithm for elliptical head contour detection. Starting with the localization of head and shoulder by an improved Gaussian Mixture Model (GMM) approach, the precise head contour is obtained by making use of the Omega shape formed from the head and shoulder. Experimental results on several benchmark datasets demonstrate the superiority of the proposed idea over the state-of-the-art in both detection accuracy and processing speed, even though there are various types of severe occlusions in faces.

An Efficient Salient Feature based Histology Image Retrieval

Due to the development of digital microscopy technologies there are huge quantity of histology images are collected and stored in database. The capability to handle and access these collections of histology images is considered as a key constituent for next generation of medical systems. These types of image retrieval system are the combination of different features play major key responsibility to retrieve images. It aims at attractive the expressive authority of visual and salient features equivalent to semantically consequential queries. Most of the existing work doesn’t go behind with different combination of features to retrieve digital images along with salient features and visual feature still becomes major problem, extraction of these features also not major problem .To overcome these issue in this paper presents a improved Gaussian mixture model that routinely extracts both salient and visual representation of features from histology image database . A Gabor filtering framework is proposed, which makes the salient regions further radiant in the saliency maps to retrieve histology images with enhanced image quality results. The major objective of the work is to found the most significant visual representation of features and salient features for a specific keyword that is connected with numerous query image retrieval processes for exact matching of results. The proposed work uses an Enhanced Artificial Fish Swarm Algorithm (AFSA) method, which aims to appreciate an best possible visual-semantic matching purpose by together allowing for the different preference of the collection of query images. It is capable to discover different types of features that summarize diverse portions of the majority representative of visual patterns and salient features for every conception. Experimental results demonstrate that proposed EAFSA system for content-based histology image retrieval system, high retrieval accurateness other than existing methods.

An Object Detection Algorithm Based on Statistical Classification of Video Stream Pixels

Moving object detection and tracking based on the video stream are both challenging and very broad application prospects of research topics. This paper presents a object detection algorithm based on statistical classification of video stream pixels that can solve moving object detection in the background illumination mutations. The algorithm determines the number of points of light mutations by the statistical number of foreground pixels of the current frame. In the light mutations object detection uses relatively simple frame-difference method, otherwise it adopts the improved Gaussian mixture model method to model. Experimental results show that the algorithm can complete detecting in the light mutation quickly and accurately and has strong robustness.

Image Object Detection Algorithm Based on Improved Gaussian Mixture Model

Aiming at poor adaptability to illumination variation and single learning rate in traditional Gaussian mixture model, an improved moving object detection algorithm based on adaptive Gaussian mixture model is proposed in this paper, so as to achieve the goal of a self-adaptive background updating model. In this paper, we analyze the existed algorithms and put forward the method to make use of color histogram matching algorithm, through introduction of illumination variation factor and update-counter of model parameter and the components number with self-adaptive selection employed to adaptively adjust learning rate, in order to greatly reduce the computation time of the algorithm and improve the real-time performance. The experiment results show that the new method can effectively adapt the scene, and has more good expansibility, robustness and stability than traditional Gaussian mixture model.

Improved Gaussian Mixture Model in Video Motion Detection

As the classical Gaussian mixture model has some problems of not considering it self’s matching degree of Gaussian density functions, model updating and the background in real video motion detection, made improvements on the three aspects. Optimized Gaussian mixture model’s overall architecture and proposed an improved algorithm according to the analysis of the definition and disadvantages of classical Gaussian mixture model. Finally, through detailed experiment, the result showed: the improved Gaussian mixture model is faster in model convergence rate in video motion detection, can quickly adapt to the changes of background and greatly decreases the fall-out ratio

Moving Objects Detection Algorithm Based on Three-Dimensional Gaussian Mixture Codebook Model Using XYZ Color Model

In order to reduce the difficulty of adjusting parameters for the codebook model and the computational complexity of probability distribution for the Gaussian mixture model in intelligent visual surveillance, a moving objects detection algorithm based on three-dimensional Gaussian mixture codebook model using XYZ color model is proposed. In this algorithm, a codebook model based on XYZ color model is built, and then the Gaussian model based on X, Y and Z components in codewords is established respectively. In this way, the characteristic of the three-dimensional Gaussian mixture model for the codebook model is obtained. The experimental results show that the proposed algorithm can attain higher real-time capability and its average frame rate is about 16.7 frames per second, while it is about 8.3 frames per second for the iGMM (improved Gaussian mixture model) algorithm, about 6.1 frames per second for the BM (Bayes model) algorithm, about 12.5 frames per second for the GCBM (Gaussian-based codebook model) algorithm, and about 8.5 frames per second for the CBM (codebook model) algorithm in the comparative experiments. Furthermore the proposed algorithm can obtain better detection quantity.

Moving Target Detection Based on the Improved Gaussian Mixture Model Background Difference Method

Gaussian mixture model background difference method is an effective method to achieve the moving target detection. According to its deficiencies of accuracy, speed and other aspects, this paper presents an improved Gaussian mixture model background difference method. Firstly, use three-frame difference method to detect the alterant area rapidly by the advantages of accuracy and fast speed. Then, use the area method to judge the results, and determine whether it is need for target extraction of the current frame by Gaussian mixture model background method, which can reduce the time of object detecting and background modeling. Meanwhile, the update strategies of the Gaussian mixture model background is improved, which can further enhance the detective accuracy and speed for the large and slow moving targets.

Moving Object Detection and Tracking in outside Environments

Background subtraction is a typical method for moving objects detection. The Gaussian mixture model is one of widely used method to model the background. However, in challenge environments, quick lighting changes, noises and shake of background can influence the detection of moving objects significantly. To solve this problem, an improved Gaussian Mixture Model is proposed in this paper. In the proposed algorithm, Objects are divided into three categories, foreground, background and middle-ground. The proposed algorithm is a segmented process. Moving objects including foreground and middle-ground are extracted firstly; then foreground is segmented from middle-ground. In this way almost middle-ground are filtered, so we can obtain a clear foreground objects. Experimental results show that the proposed algorithm can detect moving objects much more precisely, and it is robust to lighting changes and shadows.

A Moving Objects Detection Method Base on Improved GMM

Motion detection is the first and important step in many computer vision applications. Gaussian mixture model is an effective way for moving objects detection, but there are some shortcomings of this model such as slow updating rate and false detection in complex background. In this paper, we proposed an improved Gaussian mixture model method. A matching distance is defined to compute the learning rate when updating the models, and we also use dual threshold to improve the matching mechanism. Experimental results show that this method can get a faster adaptation to background and better contour of the moving objects.