• Journal of Internet Computing and Services
• /
• v.5 no.2
• /
• pp.69-73
• /
• 2004

A speaker model In speaker recognition system is to be trained from a large data set gathered in multiple sessions. Large data set requires large amount of memory and computation, and moreover it's practically hard to make users utter the data inseveral sessions. Recently the incremental adaptation methods are proposed to cover the problems, However, the data set gathered from multiple sessions is vulnerable to the outliers from the irregular utterance variations and the presence of noise, which result in inaccurate speaker model. In this paper, we propose an incremental robust adaptation method to minimize the influence of outliers on Gaussian Mixture Madel based speaker model. The robust adaptation is obtained from an incremental version of M-estimation. Speaker model is initially trained from small amount of data and it is adapted recursively with the data available in each session, Experimental results from the data set gathered over seven months show that the proposed method is robust against outliers.

• Proceedings of the Acoustical Society of Korea Conference
• /
• 1993.06a
• /
• pp.249-252
• /
• 1993

In speech signal processing, The accurate detection of the voiced/unvoiced is important for robust word recognition and analysis. This algorithm is based on the MD in the frame of speech signals that does not require statistical information about either signal or background-noise to decide a voiced/unvoiced. This paper presents a method of estimation the Characteristic of Magnitude Distribution from noisy speech and also of estimation the optimal threshold based on the MD of the voiced/unvoiced decision. The performances of this detectors is evaluated and compared to that obtained from classifying other paper.

• Proceedings of the KIPE Conference
• /
• 1997.07a
• /
• pp.332-335
• /
• 1997

In the industrial motor drive system, a shift torsional vibration si often generated when a motor and a load are connected with a flexible shaft. This paper treats the vibration suppression control of such a system. In this paper, the state feedback controller of the two-mass resonant system using the H$\infty$ filter is proposed. The H$\infty$ filter is robust in noise and disturbance. Simulation results show the validity proposed controller.

• 제어로봇시스템학회:학술대회논문집
• /
• 1991.10b
• /
• pp.1730-1733
• /
• 1991

In this paper, we two dimensional processing method is presented, which can accurately determine the orientation of an part. Matching bewteen the object and the object contour is decomposed to estimate the orientation of the object and to evaluate the similarity new approach is very robust with respect to noise and no preprocessing of the contour is required. Also, this method has many advantages over the converntional correlation technique. With only a few uniformly sampled points, this method can estimate the accurate orientation in an efficient manner even in a noisy environment.

• Journal of the Korean Institute of Telematics and Electronics B
• /
• v.33B no.3
• /
• pp.120-125
• /
• 1996

A new hit-and-miss ratio transform is introduced as a modified hit-and-miss transform to be robust to noise, which uses a quasi-matching technique based on the fitting ratio functions. And a new gray-level object segmentation algorithm is proposed, which is based on the hit-and-miss ratio transform and threshold decomposition. The proposed segmentation images, and is similarly applicable to segmentation of an object with specific shapes form natural real images.

• ETRI Journal
• /
• v.30 no.6
• /
• pp.865-867
• /
• 2008

A conventional environment adaptation for robust speech recognition is usually conducted using transform-based techniques. Here, we present a discriminative adaptation strategy based on a multi-condition-trained model, and propose a new method to provide universal application to a new environment using the environment's specific conditions. Experimental results show that a speech recognition system adapted using the proposed method works successfully for other conditions as well as for those of the new environment.

• ETRI Journal
• /
• v.30 no.5
• /
• pp.753-755
• /
• 2008

In this letter, we propose a new histogram equalization technique for feature compensation in speech recognition under noisy environments. The proposed approach combines a signal-to-noise-ratio-dependent feature reconstruction method and the class histogram equalization technique to effectively reduce the acoustic mismatch present in noisy speech features. Experimental results from the Aurora 2 task confirm the superiority of the proposed approach for acoustic feature compensation.

• Proceedings of the IEEK Conference
• /
• 2008.06a
• /
• pp.1021-1022
• /
• 2008

This paper proposes an object segmentation method using centre of mass model and $CIELa^*b^*$ color feature. The proposed method detects moving objects using geometric and colorimetic information. The method is robust to illumination changes and it reduces noise by block-wise computation.

• 제어로봇시스템학회:학술대회논문집
• /
• 1990.10b
• /
• pp.1196-1199
• /
• 1990

The authors proposed a new two-dimensional(2D) positioning system by use of M-array suitable for noisy environment in '88KACC and its revised version in '89KACC. This 2D positioning system is further improved to be used in practice; the computation time is improved by use of vector signal processor and the focussing process is improved by use of an electrically controlled zoom lense. It is shown that this system is robust to noise and also to misalignment of devices.

• Proceedings of the IEEK Conference
• /
• 1999.11a
• /
• pp.507-510
• /
• 1999

Inspection of BGAs presents several challenges for modem measurement equipment. No only must these systems be fast and accurate, they must deal with the special challenges presented by very small shiny metal spheres. For accurate measurement, we propose an algorithm which fits for estimating the accurate ball height using 2-D curve-fitting algorithm. The real boundary between two adjacent pixels and the real ball diameter are measured with subpixel accuracy Experimental results show that the proposed method calculates the ball height and diameter with subpixel accuracy and is robust in local noise with low measurement error.