• The Journal of the Acoustical Society of Korea
• /
• v.15 no.6
• /
• pp.90-98
• /
• 1996

In this paper we have proposed the sign language transformation system for deaf based on a morpheme analysis. The proposed system extracts phoneme components and connection informations of the input character sequence by using a morpheme analysis. And then the sign image obtained by component analysis is correctly and automatically generated through the sign image database. For the effective sign language transformation, the language description dictionary which consists of a morpheme analysis part for analysis of input character sequence and sign language description part for reference of sign language pattern is costructed. To avoid the duplicating sign language pattern, the pattern is classified a basic, a compound and a similar sign word. The computer simulation shows the usefulness of the proposed system.

• Journal of the Korean Institute of Telematics and Electronics B
• /
• v.31B no.5
• /
• pp.117-122
• /
• 1994

Well-trained neural networks have low sensitivity to input errors. Also, the sensitivity to weigh errors must be considered when implementing neural networks with hardware of limited precision. In this paper, we derive the sensitivity of the Madaline to weight perturbation or input errors in terms of the trained weights, the input pattern, and the variance of weight perturbation or the probability of input errors. The result is verified with a simulation of the Madaline recognizing handwritten digits.

• The Journal of the Korea Contents Association
• /
• v.10 no.4
• /
• pp.10-18
• /
• 2010

This paper presents a optimization of sigmoid activation function parameter using genetic algorithms and pattern recognition analysis in input space of two spirals benchmark problem. To experiment, cascade correlation learning algorithm is used. In the first experiment, normal sigmoid activation function is used to analyze the pattern classification in input space of the two spirals problem. In the second experiment, sigmoid activation functions using different fixed values of the parameters are composed of 8 pools. In the third experiment, displacement of the sigmoid function to determine the value of the three parameters is obtained using genetic algorithms. The parameter values applied to the sigmoid activation functions for candidate neurons are used. To evaluate the performance of these algorithms, each step of the training input pattern classification shows the shape of the two spirals.

• Journal of the Korea Institute of Information Security & Cryptology
• /
• v.21 no.4
• /
• pp.47-60
• /
• 2011

With the increase in the use of mobile banking service, mobile banking has become an attractive target to attackers. Even though many security measures are applied to the current mobile banking service, some threats such as physical theft or penetration to a mobile device from remote side are still remained as unsolved. With aiming to fill this void, we propose a novel approach to prevent e-financial incidents by analyzing mobile device user's input patterns. This approach helps us to distinguish between original user's usage and attacker's usage through analyzing personal input patterns such as input time-interval, finger pressure level on the touch screen. Our proposed method shows high accuracy, and is effective to prevent the e-finance incidents proactively.

• Journal of IKEEE
• /
• v.23 no.1
• /
• pp.164-172
• /
• 2019

Current clinical methods for diagnosing urination disorder are invasive, expensive, and very inconvenient to perform continuous monitoring. EIT is a non-invasive technique that injects electrical current through an external electrodes and measures the induced voltage to visualize the internal electrical (impedance) characteristics, which makes it possible to monitor bladder conditions with low cost. The signal characteristics of the measured voltage data changes according to the current pattern injected through the electrode and affects reconstruction performance. In this paper, image reconstruction performance is compared and analyzed according to the injected current patterns to maximize the sensitivity to the variation of bladder size.

• Journal of Institute of Control, Robotics and Systems
• /
• v.9 no.9
• /
• pp.720-726
• /
• 2003

A virtual keyboard may be efficient as a new mobile input device supporting QWERTY keyboard layout. As a preliminary study for developing a virtual keyboard, the typing pattern of a skilled human is investigated. In the study the touch-positions of the fingers are measured with a touchscreen while five skilled typists perform typing of long sentences. From these measurements it can be observed that the groups of touch-positions are classified into alphabetic characters. Though there are some mismatches, we can find constant distances capable of being discriminated among the groups. Based on the analysis the prediction algorithm of the constant distance is proposed and evaluated, which is useful for realization of a portable virtual keyboard.

• Journal of the Korea Institute of Information and Communication Engineering
• /
• v.20 no.1
• /
• pp.215-221
• /
• 2016

In this thesis we introduce a large quantity of input interface system using a low cost single chip microcontroller which is consists of walking board with 1600 switches, RS485 communication for switch data communication and PC application software for walking pattern analysis. When a pedestrian walks on the walking board, the pattern analysis of foot pressed switches can be utilized on diverse divisions of sports and industry such as walking physical therapy, dancing, a large quantity of sensors interface system, etc.

• Journal of the Institute of Electronics Engineers of Korea CI
• /
• v.45 no.5
• /
• pp.56-62
• /
• 2008

This paper proposes the use of decision tree classification for the measuring pattern recognition from industrial Computed Radiography(CR) images used in nondestructive evaluation(NDE) of steel-tubes. It appears that NDE problems are naturally desired to have machine learning techniques identify patterns and their classification. The attributes of decision tree are taken from NDE test procedure. Geometric features, such as radiative angle, gradient and distance, are estimated from the analysis of input image data. These factors are used to make it easy and accurate to classify an input object to one of the pre-specified classes on decision tree. This algerian is to simplify the characterization of NDE results and to facilitate the determination of features. The experimental results verify the usefulness of proposed algorithm.

• ETRI Journal
• /
• v.36 no.6
• /
• pp.979-987
• /
• 2014

This paper describes new tactile feedback patterns and the effect of their input performance for a button GUI activated by a tap gesture on mobile touch devices. Based on an analysis of touch interaction and informal user tests, several tactile feedback patterns were designed. Using these patterns, three user experiments were performed to investigate appropriate tactile feedback patterns and their input performance during interaction with a touch button. The results showed that a tactile pattern responding to each touch and release gesture with a rapid response time and short falling time provides the feeling of physically clicking a button. The suggested tactile feedback pattern has a significantly positive effect on the number of typing errors and typing task completion time compared to the performance when no feedback is provided.

• The Transactions of The Korean Institute of Electrical Engineers
• /
• v.67 no.1
• /
• pp.114-123
• /
• 2018

This paper concerns a design and learning method of softmax function neural networks based on K-means clustering. The partial discharge data Information is preliminarily processed through simulation using an Epoxy Mica Coupling sensor and an internal Phase Resolved Partial Discharge Analysis algorithm. The obtained information is processed according to the characteristics of the pattern using a Motor Insulation Monitoring System program. At this time, the processed data are total 4 types that void discharge, corona discharge, surface discharge and slot discharge. The partial discharge data with high dimensional input variables are secondarily processed by principal component analysis method and reduced with keeping the characteristics of pattern as low dimensional input variables. And therefore, the pattern classifier processing speed exhibits improved effects. In addition, in the process of extracting the partial discharge data through the MIMS program, the magnitude of amplitude is divided into the maximum value and the average value, and two pattern characteristics are set and compared and analyzed. In the first half of the proposed partial discharge pattern classifier, the input and hidden layers are classified by using the K-means clustering method and the output of the hidden layer is obtained. In the latter part, the cross entropy error function is used for parameter learning between the hidden layer and the output layer. The final output layer is output as a normalized probability value between 0 and 1 using the softmax function. The advantage of using the softmax function is that it allows access and application of multiple class problems and stochastic interpretation. First of all, there is an advantage that one output value affects the remaining output value and its accompanying learning is accelerated. Also, to solve the overfitting problem, L2-normalization is applied. To prove the superiority of the proposed pattern classifier, we compare and analyze the classification rate with conventional radial basis function neural networks.