• The Journal of the Acoustical Society of Korea
• /
• 제15권4E호
• /
• pp.28-36
• /
• 1996

In this paper, we propose a new evaluation methodology which can more clearly show the performance of the allophone modeling algorithm generally used in large vocabulary speech recognition. The proposed evaluation method shows the running characteristics and limitations of the modeling algorithm by testing how the variation of phonetic environments of training data affects the recognition performance and the desirable number of free parameters to be estimated. Using the method, we experiment results, we conclude that, in vocabulary-independent recognition task, the phonetic diversity of training data greatly affects the robustness of model, and it is necessary to develop a proper measure which can determine the number of states compromizing the robustness and the precision of the HMnet better than the conventional modeling efficiency.

• 전자공학회논문지B
• /
• 제30B권2호
• /
• pp.69-77
• /
• 1993

Automatic image recognition system is essential for a better man-to machine interaction. Because of the noise and deformation due to the sensor operation, it is not simple to build an image recognition system even for the fixed images. In this paper neural network which has been reported to be adequate for pattern recognition task is applied to the fixed and variational(rotation, size, position variation for the fixed image)recognition with a hope that the problems of conventional pattern recognition techniques are overcome. At fixed image recognition system. ART model is trained with face images obtained by camera. When recognizing an matching score. In the test when wigilance level 0.6 - 0.8 the system has achievel 100% correct face recognition rate. In the variational image recognition system, 65 invariant moment features sets are taken from thirteen persons. 39 data are taken to train multi-layer perceptron and other 26 data used for testing. The result shows 92.5% recognition rate.

• Journal of Advanced Marine Engineering and Technology
• /
• 제17권4호
• /
• pp.1-10
• /
• 1993

This paper introduces the outline of hull structure to the sorts of LNG carrier briefly. Especially, explains in detail for the insulation system of Moss Rosenberg Verft spherical tank type LNG carrier. It is not easy task to calculate exactly the temperature distribution of hull because of very complicated structure of hull. Therefore, in this paper by the adequate modeling of the Moss Rosengerg spherical tank type LNG carrier, a program is developed which calculate the temperature distribution of every hull and estimate the heat influx from every hull and output the BOR according to the variation of atmospheric conditions on boyage.

• Ocean and Polar Research
• /
• 제23권2호
• /
• pp.161-172
• /
• 2001

A model-to-model comparison is attempted between Princeton Ocean Model (POM) and Miami Isopycnic Coordinate Ocean Model (MICOM) as a first step to extend our knowledge of models' performances in studying the East Sea circulation. The two models have fundamentally different numerical schemes and boundary conditions imposed on these models are not exactly the same each other. This study indicates that MICOM has a critical weak point in that it does not reproduce the shallow surface currents properly while it handles the thermohaline processes and associated movements of intermediate and deep waters efficiently. It is suggested that the mixed layer scheme needs to be modified so that it can match with inflow boundary conditions in order to reproduce the surface currents properly in MICOM. POM reproduces the surface current pattern better than MICOM, although the surface currents in POM appear to undergo the unrealistic seasonal variation and have exaggeratedly large vertical scale. These defects seem to arise during the process of adapting POM to the East Sea, and removing these defects is left as a future task.

• 한국산업융합학회 논문집
• /
• 제11권4호
• /
• pp.171-175
• /
• 2008

The bigger size of the wafer to the meet the need of higher productivity and the smaller digital products for convene are examples of the main trends of the market among others. The higher integrity of chips is also an important task in semiconductor industry. In this thesis the COF(chop on film) technology has reviewed to investigate the best combinations of required functions and to develop a new system to save the installation area and to adapt to the variation of layout with flexibility. The new system shows the better and provides the improved safety with smaller size.

• 대한전기학회:학술대회논문집
• /
• 대한전기학회 2000년도 하계학술대회 논문집 D
• /
• pp.2389-2391
• /
• 2000

To design high quality products at low cost is one of very important tasks for engineers. Design optimization for performances can be one solution in this task. There is the robust design which has been proved effectively in many fields of engineering design. In this paper, the concept of robust design is introduced and combined to the fuzzy optimization method and the fuzzy logic system method with non-singleton. These methods are applied for data analysis to get optimum parameters and to reduce experiments. The optimum parameter set points are obtained by the proposed methods. These methods are applied to a filter circuit, a part of the audio circuit of mobile radio transceiver. The simulation results are compared each other. The new methods reduce and predict the effect of parameter variation sources

• 대한기계학회논문집A
• /
• 제21권6호
• /
• pp.981-987
• /
• 1997

Modern helicopter rotor blades with non-homogeneous cross section composed of anisotropic material rquire highly sophisticated structural analysis. Variation in cross section geometry makes this task of analysis more complicated. Since rotor blades generally are much longer than their lateral dimensions, one-dimensional models seem feasible, at least from a computational point of view. Therefore determination of the principal coordinate system is very important to remove the structural coupling for one-dimensional beam modelling. In this study, shear center, and principal direction. The method will be verified by comparing the results with confirmed experimental results.

• International Journal of Safety
• /
• 제10권1호
• /
• pp.32-35
• /
• 2011

Lifting capacity and difficulty of task are influenced by body posture. In RULA and REBA, the body was divided into segments which formed two groups, A and B. Group A includes the upper and lower arm and wrist while group B includes the neck, trunk and legs. This ensures that whole body posture is recorded so that any awkward or constrained posture of the legs, trunk or neck which might influence the posture of the upper limb. This study aimed to measure MVC (maximum voluntary contraction) and subjective judgment in psychophysical method (Borg's scale) according to trunk and upper arm angle and to analyze results statistically. The results of this study were that lifting capacity was more influenced by interaction of body posture rather than angles of each part, and MVC variation according to trunk and upper arms angles should different patterns. This means that we consider the interaction of trunk angles and upper arm angles when we access risk factors of the postures. This survey would be also the basic data to evaluate difficulty of lifting tasks according to body postures ergonomically.

• 한국정보처리학회:학술대회논문집
• /
• 한국정보처리학회 2019년도 춘계학술발표대회
• /
• pp.525-528
• /
• 2019

Atmospheric particle removal is a challenging task and attacks wide interests in computer vision filed. In this paper, we proposed a single image snow removal framework based on deep residual networks. According to the fact that there are various snow sizes in a snow image, the inception module which consists of different filter kernels was adopted to extract multiple resolution features of the input snow image. Except the traditional mean square error loss, the perceptual loss and total variation loss were employed to generate more clean images. Experimental results on synthetic and realistic snow images indicated that the proposed method achieves superior performance in respect of visual perception and objective evaluation.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• 제13권2호
• /
• pp.871-891
• /
• 2019

Visual tracking is a challenging task that needs learning an effective model to handle the changes of target appearance caused by factors such as pose variation, illumination change, occlusion and motion blur. In this paper, a novel tracking algorithm based on weighted local sub-space reconstruction error is presented. First, accounting for the appearance changes in the tracking process, a generative weight calculation method based on structural reconstruction error is proposed. Furthermore, a template update scheme of occlusion-aware is introduced, in which we reconstruct a new template instead of simply exploiting the best observation for template update. The effectiveness and feasibility of the proposed algorithm are verified by comparing it with some state-of-the-art algorithms quantitatively and qualitatively.