GMM Algorithm Research Articles

A distributed incremental update scheme for probability distribution of wind power forecast error

Due to the uncertainty of distributed wind generations (DWGs), a better understanding of the probability distributions (PDs) of their wind power forecast errors (WPFEs) can help market participants (MPs) who own DWGs perform better during trading. Under the premise of an accurate PD model, considering the correlation among DWGs and absorbing the new information carried by the latest data are two ways to maintain an accurate PD. These two ways both require the historical and latest wind power and forecast data of all DWGs. Each MP, however, only has access to the data of its own DWGs and may refuse to share these data with MPs belonging to other stakeholders. Besides, because of the endless generation of new data, the PD updating burden increases sharply. Therefore, a distributed strategy is used to avoid raw data collection. In addition, the incremental learning strategy is further applied to reduce the updating burden. Finally, a distributed incremental update scheme is proposed to make each MP continually acquire the latest conditional PD of its DWGs’ WPFE. Specifically, the Gaussian-mixture-model-based (GMM-based) joint PD is first used to characterize the correlation among DWGs. Then, a distributed modified incremental GMM algorithm is proposed to enable MPs to update the parameters of the joint PD in a distributed and incremental manner. After that, a distributed derivation algorithm is further proposed to make MPs derive their conditional PD of WPFE from the joint one in a distributed way. Combining the two original algorithms, the complete distributed incremental update scheme is finally achieved, by which each MP can continually obtain its latest conditional PD of its DWGs’ WPFE via neighborhood communication and local calculation with its own data. The effectiveness, correctness, and efficiency of the proposed scheme are verified using the dataset from the NREL.

Efficient Particle MCMC with GMM Likelihood Representation

Particle Markov Chain Monte Carlo (PMCMC) is a widely used method to handle estimation problem in the context of nonlinear structural dynamic models whose likelihood function is analytically intractable. PMCMC can be constructed upon a GMM likelihood representation when one does not want to rely on the structural form of the measurement equation (Gallant et al 2016). It only requires to compute moment conditions available from the structural model. However, particle filter with GMM may suffer from high degeneracy of particle weights which severely affects the accuracy of Monte Carlo approximations and in turn Markov Chain Monte Carlo estimates. This work is concerned with revising particle GMM algorithm as proposed in Gallant et al in order to reduce the depletion problem. Estimation results of stochastic volatility models show that the efficient block sampling strategy as proposed in Doucet et al (2006) can outperform particle GMM and in turn deliver more reliable MCMC estimates. Auxiliary particle filter (Doucet et al, 2011) is also proposed as an alternative strategy to the block sampling approach. However, in the intended experiments it does not seem to be very effective. Thus some of the assumptions needed to estimate structural nonlinear state space models can be weakened and requiring only available moment conditions without affecting dramatically the conclusions.

Lameness detection of dairy cows based on a double normal background statistical model

Dairy cow lameness is one of the diseases that affect the health of dairy cows, and the rapid and accurate detection of dairy cow lameness is an important research field for modern dairy cow farming. In this study, a lameness detection method based on a double normal distribution statistical model was proposed. The combination of these features could be used to determine the lameness of dairy cows. To verify the validity of the lameness detection algorithm proposed in this research, the detection accuracy (D-Accuracy), detection accuracy of lame (L-Accuracy), and detection accuracy of non-lame cows (N-Accuracy) were used as indices to evaluate the algorithm performance. The results showed that the total correct rate of D-Accuracy was 93.75%, that of L-Accuracy was 90.00%, and that of N-Accuracy was 100.00%. Compared with the typical GMM algorithm, the proposed algorithm was more suitable for detecting dairy cow lameness. The results also show that the robustness against the environment was stronger and that the false positive rate of the target dairy cows was reduced by 18.71%. This research has important significance for the early detection and treatment of dairy cow lameness, which could be used to keep production level high and guarantee good animal welfare.

Efficient Particle MCMC with GMM Likelihood Representation

Particle Markov Chain Monte Carlo (PMCMC) is a widely used method to handle estimation problem in the context of nonlinear structural dynamic models whose likelihood function is analytically intractable. PMCMC can be constructed upon a GMM likelihood representation when one does not want to rely on the structural form of the measurement equation (Gallant et al 2016). It only requires to compute moment conditions available from the structural model. However, particle filter with GMM may suffer from high degeneracy of particle weights which severely affects the accuracy of Monte Carlo approximations and in turn Markov Chain Monte Carlo estimates. This work is concerned with revising particle GMM algorithm as proposed in Gallant et al in order to reduce the depletion problem. Estimation results of stochastic volatility models show that the efficient block sampling strategy as proposed in Doucet et al (2006) can outperform particle GMM and in turn deliver more reliable MCMC estimates. Auxiliary particle filter (Doucet et al, 2011) is also proposed as an alternative strategy to the block sampling approach. However, in the intended experiments it does not seem to be very effective. Thus some of the assumptions needed to estimate structural nonlinear state space models can be weakened and requiring only available moment conditions without affecting dramatically the conclusions.

Design of Moving Object Detector Based on Modified GMM Algorithm for UAV Collision Avoidance

Design of Moving Object Detector Based on Modified GMM Algorithm for UAV Collision Avoidance

Moving Target Detection Algorithm for Forest Fire Smoke Recognition with Improved ViBe

In order to discover the forest fire at the early stage, the video based fire smoke detection system should be developed. This paper proposes an improved ViBe algorithm to overcome the drawbacks of ViBe which is widely used in forest fire smoke recognition. To make the algorithm faster, the improved ViBe algorithm introduces a switch variable to control the model updating every 2 frames. Also, the new ViBe algorithm contains a method to detect if the scene is changing, and upon finding the changing, a fast updating strategy would be used to increase the adaptive ability. The experiments are done from 2 aspects, to test the time consuming and adaptive ability of the improved algorithm. The experiments results show that the improved ViBe algorithm has better performance compared to ViBe algorithm and GMM algorithm, and can detect the forest fire smoke accurately in different conditions.

Research on Moving Target Recognition for Vehicle Driving Robot with Remote Operation Based on OpenCV

Owing to the need of vehicle driving with remote operation and driverless technology, it has become a hot topic to deal with the real environment returned by the camera placed in the vehicle which serves as a sensor. In order to realize the recognition of moving objects, such as motor vehicles in the returned data stream, the C++ interface of OpenCV can be employed to program and detect the dynamic target by using GMM algorithm, achieving the effect of real-time detection of the environment picture returned by the camera and marking moving targets such as vehicles.

Multi-model multivariate Gaussian process modelling with correlated noises

A composite multiple-model approach based on multivariate Gaussian process regression (MGPR) with correlated noises is proposed in this paper. In complex industrial processes, observation noises of multiple response variables can be correlated with each other and process is nonlinear. In order to model the multivariate nonlinear processes with correlated noises, a dependent multivariate Gaussian process regression (DMGPR) model is developed in this paper. The covariance functions of this DMGPR model are formulated by considering the “between-data” correlation, the “between-output” correlation, and the correlation between noise variables. Further, owing to the complexity of nonlinear systems as well as possible multiple-mode operation of the industrial processes, to improve the performance of the proposed DMGPR model, this paper proposes a composite multiple-model DMGPR approach based on the Gaussian Mixture Model algorithm (GMM-DMGPR). The proposed modelling approach utilizes the weights of all the samples belonging to each sub-DMGPR model which are evaluated by utilizing the GMM algorithm when estimating model parameters through expectation and maximization (EM) algorithm. The effectiveness of the proposed GMM-DMGPR approach is demonstrated by two numerical examples and a three-level drawing process of Carbon fiber production.

배경 모델 갱신을 통한 코드북 기반의 전배경 분할

최근 코드북 기반의 전 배경 분리 알고리즘에 대한 연구가 활발히 진행되고 있다. 코드북은 입력 영상 시퀀스로부터 화소당 하나씩 만들어 지는데, 코드북 내의 각 코드워드는 동일 위치의 훈련 화소들을 대상으로 양자화를 수행한 클러스터 대표 벡터이다. 일반적인 코드북 기반 방법들은 초기 배경 모델 생성을 위하여 긴 시간동안 훈련 샘플들의 학습 과정을 거친다. 본 논문에서는 초기 몇 장의 프레임으로 부터 간단한 중위수 연산을 통하여 초기 배경 모델을 생성하고, 시간의 흐름에 따라 변화된 배경 정보를 포함할 수 있도록 코드워드의 사용 빈도수에 기반하여 배경 모델을 갱신한다. 제안한 알고리즘을 OpenCV 3.0과 연동하여 C언어로 구현하여 몇 개의 PETS2009 데이터에 적용 실험하였다. 해당 데이터는 준-주기적 움직임을 갖는 영상 시퀀스, 이동 물체의 일시 정지 등의 시나리오를 포함하고 있다. 실험을 통하여 제안한 방식이 GMM 알고리즘, 표준 코드북 알고리즘에 비하여 우수한 성능을 나타내고 있음을 확인하였다. Recently, a foreground-background segmentation using codebook model has been researched actively. The codebook is created one for each pixel in the image. The codewords are vector-quantized representative values of same positional training samples from the input image sequences. The training is necessary for a long time in the most of codebook-based algorithms. In this paper, the initial codebook model is generated simply using median operation with several image frames. The initial codebook is updated to adapt the dynamic changes of backgrounds based on the frequencies of codewords that matched to input pixel during the detection process. We implemented the proposed algorithm in the environment of visual c++ with opencv 3.0, and tested to some of the public video sequences from PETS2009. The test sequences contain the various scenarios including quasi-periodic motion images, loitering objects in the local area for a short time, etc. The experimental results show that the proposed algorithm has good performance compared to the GMM algorithm and standard codebook algorithm.

Vehicles detection in complex urban traffic scenes using Gaussian mixture model with confidence measurement

Aiming to efficiently resolve the problem that the subtraction background model is easily contaminated by slow-moving or temporarily stopped vehicles, the Gaussian mixture model with confidence measurement (GMMCM) is proposed for vehicle detection in complex urban traffic scenes. According to the current traffic state, each pixel of background model is set a CM. Whether to update the background model and the corresponding adaptive learning rate depends on if the current pixel point is in confidence period. Using the real-world urban traffic videos, the first experiments are conducted by GMMCM, compared with three commonly used models including GMM, self-adaptive GMM (SAGMM) and local parameter learning algorithm for the GMM (LPLGMM). The first experimental results show that GMMCM excels GMM, SAGMM and LPLGMM in keeping the background model being unpolluted from slow-moving or temporarily stopped vehicles. The second experiments are conducted by GMMCM, compared with visual background extractor, sigma-delta with CM, SAGMM, LPLGMM and GMM. The average recalls of six methods are 0.899, 0.753, 0.679, 0.420, 0.447 and 0.205, and the average F-measures of six methods are 0.636, 0.612, 0.592, 0.373, 0.330 and 0.179, respectively. All experimental results show the effectiveness of the proposed GMMCM in vehicles detection of complex urban traffic scenes.

Quality-related process monitoring for dynamic non-Gaussian batch process with multi-phase using a new data-driven method

In this paper, a quality-related monitoring scheme of batch process using multi-phase dynamic non-Gaussian model is presented. Product quality of a batch process is difficult to be effectively guaranteed because of its frequent start-stop operation, variable operating conditions, strong dynamic and non-Gaussian character of process data. A direct dynamic PLS (DDPLS), in which weighted time-lagged matrix is used to extract dynamic components, is introduced to the dynamic problem. Meanwhile, independent component analysis (ICA) is proposed to deal with non-Gaussianity of dynamic components in DDPLS. Considering most batch processes are multi-phase in nature, in order to well describe the characteristics of every phase and set up sub-models, GMM algorithm is adopted for phase division and fuzzy membership method for transition identification. TE benchmark is used to verify the validity and superiority of our new method over traditional PLS, DPLS. Then the new method is applied to a real hot strip mill production plant.

An Application of GMM Algorithm in Signature Skew Detection

An Application of GMM Algorithm in Signature Skew Detection

Research on moving object detection based on improved mixture Gaussian model

For the deficiency of the mixture Gaussian model (GMM), an improved GMM algorithm is proposed, which can be applied to moving object detection. Combined with codebook detection algorithm, the GMM model is initialized by video images pixels statistical mean and variance, and updated model using parameters confidence interval. The experiment results indicate that the proposed update model can detect moving objects in complex background effectively and has good robustness.

연속적인 배경 모델 학습을 이용한 코드북 기반의 전경 추출 알고리즘

Detection of moving objects is a fundamental task in most of the computer vision applications, such as video surveillance, activity recognition and human motion analysis. This is a difficult task due to many challenges in realistic scenarios which include irregular motion in background, illumination changes, objects cast shadows, changes in scene geometry and noise, etc. In this paper, we propose an foreground extraction algorithm based on codebook, a database of information about background pixel obtained from input image sequence. Initially, we suppose a first frame as a background image and calculate difference between next input image and it to detect moving objects. The resulting difference image may contain noises as well as pure moving objects. Second, we investigate a codebook with color and brightness of a foreground pixel in the difference image. If it is matched, it is decided as a fault detected pixel and deleted from foreground. Finally, a background image is updated to process next input frame iteratively. Some pixels are estimated by input image if they are detected as background pixels. The others are duplicated from the previous background image. We apply out algorithm to PETS2009 data and compare the results with those of GMM and standard codebook algorithms.

Study of empirical mode decomposition and spectral analysis for stress and emotion classification in natural speech

Two new approaches to the feature extraction process for automatic stress and emotion classification in speech are proposed and examined. The first method uses the empirical model decomposition (EMD) of speech into intrinsic mode functions (IMF) and calculates the average Renyi entropy for the IMF channels. The second method calculates the average spectral energy in the sub-bands of speech spectrograms and can be enhanced by anisotropic diffusion filtering of spectrograms. In the second method, three types of sub-bands were examined: critical, Bark and ERB. The performance of the new features was compared with the conventional mel frequency cepstral coefficients (MFCC) method. The modeling and classification process applied the classical GMM and KNN algorithms. The experiments used two databases containing natural speech, SUSAS (annotated with three different levels of stress) and the Oregon Research Institute (ORI) data (annotated with five different emotions: neutral, angry, anxious, dysphoric, and happy). For the SUSAS data, the best average recognition rates of 77% were obtained when using spectrogram features calculated within ERB bands and combined with anisotropic filtering. For the ORI data, the best result of 53% was obtained with the same method but without anisotropic filtering. Both the GMM and KNN classifiers showed similar performance. These results indicated that the spectrogram patterns provide promising results in the case of stress recognition, however further improvements are needed in the case of emotion recognition.