• IEIE Transactions on Smart Processing and Computing
• /
• v.2 no.6
• /
• pp.332-338
• /
• 2013

We propose a method for target detection in Infrared images. In order to effectively detect a target region from an image with noises and clutters, spatial information of the target is first considered by analyzing pixel distributions of projections in horizontal and vertical directions. These distributions are represented as Gaussian distributions, and Gaussian Mixture Model is created from these distributions in order to find thresholding points of the target region. Through analyzing the calculated Gaussian Mixture Model, the target region is detected by eliminating various backgrounds such as noises and clutters. This is performed by using a novel thresholding method which can effectively detect the target region. As experimental results, the proposed method has achieved better performance than existing methods.

• The Journal of the Korea institute of electronic communication sciences
• /
• v.7 no.4
• /
• pp.733-739
• /
• 2012

In this paper, a smoke detection algorithm robust to brightness and color variations depending on time and weather is proposed. The proposed smoke detection algorithm specifies the candidate region using difference images of input and background images, determines smoke by comparing feature coefficients of Gaussian mixture model of difference images. Thresholds for specifying candidate region is divided by four levels according to average brightness and chrominance of input images. Clusters of Gaussian mixture models of difference images are aligned according to average brightness. Smoke is determined by comparing distance of Gaussian mixture model parameters. The proposed algorithm is implemented by media dedicated DSP. As results of experiments, it is shown that the proposed algorithm is effective to detect smoke with camera installed outdoor.

• ETRI Journal
• /
• v.43 no.1
• /
• pp.74-81
• /
• 2021

The mixture model is a very powerful and flexible tool in clustering analysis. Based on the Dirichlet process and parsimonious Gaussian distribution, we propose a new nonparametric mixture framework for solving challenging clustering problems. Meanwhile, the inference of the model depends on the efficient online variational Bayesian approach, which enhances the information exchange between the whole and the part to a certain extent and applies to scalable datasets. The experiments on the scene database indicate that the novel clustering framework, when combined with a convolutional neural network for feature extraction, has meaningful advantages over other models.

• The Journal of the Acoustical Society of Korea
• /
• v.29 no.1
• /
• pp.56-61
• /
• 2010

This paper describes optimization methods of acoustic models in HMM-based continuous speech recognition. Most of the conventional speech recognition systems use the same number of Gaussian mixture components for each HMM state. However, since the number of data samples available for each state is different from each other, it is possible to reduce the overall number of model parameters and the computational cost at the decoding step by optimizing the number of Gaussian mixture components. In this study, we introduced the Gaussian mixture weight term at the merging stage of Gaussian components in the minimum description length (MDL) based acoustic modeling optimization. Experimental results showed that the proposed method can obtain better ASR accuracy than the previous optimization method which does not consider the Gaussian mixture weight term.

• Journal of KIISE:Software and Applications
• /
• v.30 no.5_6
• /
• pp.454-460
• /
• 2003

In this paper, we propose an efficient method for recognizing human lips. Based on Point Distribution Model, a lip shape is represented as a set of points. We calculate a lip model and the distribution of shape parameters using Principle Component Analysis and Gaussian mixture, respectively. The Expectation Maximization algorithm is used to determine the maximum likelihood parameter of Gaussian mixture. The lip contour model is derived by using the gray value changes at each point and in regions around the point and used to search the lip shape in a image. The experiments have been performed for many images, and show very encouraging result.

• Journal of Korea Multimedia Society
• /
• v.7 no.11
• /
• pp.1531-1539
• /
• 2004

In this paper, we propose an efficient method for locating human lips. Based on Point Distribution Model and Principle Component Analysis, a lip shape model is built. Lip boundary model is represented based on the concatenated gray level distribution model. We calculate the distribution of shape parameters using Gaussian mixture. The problem to locate lip is simplified as the minimization problem of matching object function. The Down Hill Simplex Algorithm is used for the minimization with Gaussian Mixture for setting initial condition and refining estimate of lip shape parameter, which can refrain iteration from converging to local minima. The experiments have been performed for many images, and show very encouraging result.

• International Journal of Fuzzy Logic and Intelligent Systems
• /
• v.13 no.1
• /
• pp.59-66
• /
• 2013

In this paper, we propose a probabilistic framework for model-based clustering of direction of arrival (DOA) data to obtain stable sound source localization (SSL) estimates. Model-based clustering has been shown capable of handling highly overlapped and noisy datasets, such as those involved in DOA detection. Although the Gaussian mixture model is commonly used for model-based clustering, we propose use of the von Mises mixture model as more befitting circular DOA data than a Gaussian distribution. The EM framework for the von Mises mixture model in a unit hyper sphere is degenerated for the 2D case and used as such in the proposed method. We also use a histogram of the dataset to initialize the number of clusters and the initial values of parameters, thereby saving calculation time and improving the efficiency. Experiments using simulated and real-world datasets demonstrate the performance of the proposed method.

• Smart Media Journal
• /
• v.2 no.2
• /
• pp.14-19
• /
• 2013

Recently, due to the increase of the use of scanned images, the text segmentation techniques, which play critical role to optimize the quality of the scanned images, are required to be updated and advanced. In this study, an algorithm has been developed based on the modification of Gaussian mixture model (GMM) by integrating the calculation of Gaussian detection gradient and the estimation of the number clusters. The experimental results show an efficient method for text segmentation in natural scenes such as storefronts, street signs, scanned journals and newspapers at different size, shape or color of texts in condition of lighting changes and complex background. These indicate that our model algorithm and research approach can address various issues, which are still limitations of other senior algorithms and methods.

• The Journal of the Acoustical Society of Korea
• /
• v.37 no.6
• /
• pp.506-511
• /
• 2018

This paper proposes an GMM(Gaussian Mixture Model)-DNN(Deep Neural Network) hybrid-based feature compensation method for effective speech recognition in noisy environments. In the proposed algorithm, the posterior probability for the conventional GMM-based feature compensation method is calculated using DNN. The experimental results using the Aurora 2.0 framework and database demonstrate that the proposed GMM-DNN hybrid-based feature compensation method shows more effective in Known and Unknown noisy environments compared to the GMM-based method. In particular, the experiments of the Unknown environments show 9.13 % of relative improvement in the average of WER (Word Error Rate) and considerable improvements in lower SNR (Signal to Noise Ratio) conditions such as 0 and 5 dB SNR.

• ETRI Journal
• /
• v.32 no.5
• /
• pp.795-800
• /
• 2010

Embedding a large vocabulary speech recognition system in mobile devices requires a reduced acoustic model obtained by eliminating redundant model parameters. In conventional optimization methods based on the minimum description length (MDL) criterion, a binary Gaussian tree is built at each state of a hidden Markov model by iteratively finding and merging similar mixture components. An optimal subset of the tree nodes is then selected to generate a downsized acoustic model. To obtain a better binary Gaussian tree by improving the process of finding the most similar Gaussian components, this paper proposes a new distance measure that exploits the difference in likelihood values for cases before and after two components are combined. The mixture weight of Gaussian components is also introduced in the component merging step. Experimental results show that the proposed method outperforms MDL-based optimization using either a Kullback-Leibler (KL) divergence or weighted KL divergence measure. The proposed method could also reduce the acoustic model size by 50% with less than a 1.5% increase in error rate compared to a baseline system.