• The Journal of the Acoustical Society of Korea
• /
• v.21 no.4
• /
• pp.361-373
• /
• 2002

In this paper, we propose a channel normalization method to improve the performance of CMS (cepstral mean subtraction) which is widely adopted to normalize a channel variation for speech and speaker recognition. CMS which estimates the channel effects by averaging long-term cepstrum has a weak point that the estimated channel is biased by the formants of voiced speech which include a useful speech information. The proposed Formant-broadened Cepstral Mean Subtraction (FBCMS) is based on the facts that the formants can be found easily in log spectrum which is transformed from the cepstrum by fourier transform and the formants correspond to the dominant poles of all-pole model which is usually modeled vocal tract. The FBCMS evaluates only poles to be broadened from the log spectrum without polynomial factorization and makes a formant-broadened cepstrum by broadening the bandwidths of formant poles. We can estimate the channel cepstrum effectively by averaging formant-broadened cepstral coefficients. We performed the experiments to compare FBCMS with CMS, PFCMS using 4 simulated telephone channels. In the experiment of channel estimation, we evaluated the distance cepstrum of real channel from the cepstrum of estimated channel and found that we were able to get the mean cepstrum closer to the channel cepstrum due to an softening the bias of mean cepstrum to speech. In the experiment of text-independent speaker identification, we showed the result that the proposed method was superior than the conventional CMS and comparable to the pole-filtered CMS. Consequently, we showed the proposed method was efficiently able to normalize the channel variation based on the conventional CMS.

• The KIPS Transactions:PartD
• /
• v.13D no.5 s.108
• /
• pp.709-716
• /
• 2006

As the software complexity increases, the development success rate decreases and failure rate increases exponentially. The failure rate related to the software size can be described by a growth function. Based on this phenomenon, this paper estimates the development success and completion rate using the Gompertz growth function. At first, we transformed a software size of numerically suggested $10^n$ into a logarithm and kept the data interval constantly. We tried to derive a functional relationship between the development success rate and the completion rate according to the change of logarithmic software size. However, we could not find a function which can represent this relationship. Therefore, we introduced the failure rate and the cancel rate which are inverse to the development success rate and completion rate, respectively. Then, we indicated the relation between development failure rate and cancel rate based on the change of software size, as a type of growth function. Finally, as we made the Gompertz growth function with the function which describes the cancel rate and the failure rate properly. We could express the actual data suitably. When you apply the growth function model that I suggested, you will be able to get the success rate and completion rate of particular site of software very accurately.

• Journal of the Korea Institute of Information and Communication Engineering
• /
• v.20 no.6
• /
• pp.1177-1182
• /
• 2016

The most industrial instruments for monitoring and control are occurring the extension problem and the external protocol compatibility. In this paper, we developed the boards for the industrial communication monitoring that are able to convert the protocol in various communication between devices and instruments. These are consisted the main board and several sub-board. They can have extension using the main board connection. The sub-board support the each communication method or data transfer. The main board was used the Atmega 2560 Microprocessor of AVR series, and the sub-boards are have the Atmega 256 or Atmega 128 in the AVR series. We have designed to connect the sub-board using placed the 4 RS485 serial slots in the main board. The sub-boards were developed to support the analog and digital I/O. These are able to have monitoring by CAN and Ethernet communication. The experimental results, we obtained good data transfer rate and conversion rate.

• The Journal of The Korea Institute of Intelligent Transport Systems
• /
• v.6 no.2
• /
• pp.95-106
• /
• 2007

This paper proposes an effective tempo feature extraction method for music information retrieval. The tempo information is modeled by the narrow-band temporal modulation components, which are decomposed into a modulation spectrum via joint frequency analysis. In implementation, the tempo feature is directly extracted from the modified discrete cosine transform coefficients, which is the output of partial MP3(MPEG 1 Layer 3) decoder. Then, different features are extracted from the amplitudes of modulation spectrum and applied to different music information retrieval tasks. The logarithmic scale modulation frequency coefficients are employed in automatic music emotion classification and music genre classification. The classification precision in both systems is improved significantly. The bit vectors derived from adaptive modulation spectrum is used in audio fingerprinting task That is proved to be able to achieve high robustness in this application. The experimental results in these tasks validate the effectiveness of the proposed tempo feature.

• The Korean Journal of Applied Statistics
• /
• v.23 no.3
• /
• pp.595-607
• /
• 2010

This paper addresses the problem of reducing the speckle noise in SAR images by wavelet transformation, using a non-local means(NLM) filter originated for Gaussian noise removal. Log-transformed SAR image makes multiplicative speckle noise additive. Thus, non-local means filtering and wavelet thresholding are used to reduce the additive noise, followed by an exponential transformation. NLM filter is an image denoising method that replaces each pixel by a weighted average of all the similarly pixels in the image. But the NLM filter takes an acceptable amount of time to perform the process for all possible pairs of pixels. This paper, also proposes an alternative strategy that uses the t-test more efficiently to eliminate pixel pairs that are dissimilar. Extensive simulations showed that the proposed filter outperforms many existing filters terms of quantitative measures such as PSNR and DSSIM as well as qualitative judgments of image quality and the computational time required to restore images.

• Journal of Korea Society of Industrial Information Systems
• /
• v.7 no.2
• /
• pp.74-81
• /
• 2002

In general, a multipath generator consists of a time delay generator, phase rotator, and amplitude attenuator, and is implemented mostly in an analog manner. Analog, or partially analog versions of a multipath generator is disadvantageous in that they may suffer from problems associated with component aging and adjustment, signal fidelity degradation stemming from repeated A/D and D/A conversion use of high frequency to achieve fine i.e., subsample fractional tin delays. This paper presents the design and implementation methodology of a full digital multipath generator which can be used in performance evaluations of digital terrestrial television as well as communications, receivers. In particular, an efficient practical method is proposed which can achieve both integer and fractional time delays simultaneously, without placing restrictions on the allowable system master clock frequency. The proposed algorithm lends itself to minimizing hardware implementation cost by relegating some fixed put of the computation involved to an IBM PC. The proposed multipath generator occupies only a single digital board space, and its experimental results are provided to corroborate the proposed implementation methodology.

• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.28 no.11C
• /
• pp.1098-1110
• /
• 2003

In this paper, we propose a homomorphic wavelet-based MMSE filter with edge detection to restore images degraded by speckle noise. In the proposed method, a noisy image is first transformed into logarithmic domain. Each pixel in the transformed image is then classified into flat and edge regions by applying DIP operator to the image restored by homomorphic directional MMSE filter. Each pixel in flat region is restored by homomorphic wavelet-based MMSE filter. Each pixel in edge region is restored by the weighted sum of the output of homomorphic wavelet-based MMSE filtering and that of homomorphic directional MMSE filtering. The restored image in spatial domain is finally obtained by applying the exponential function to the restored image in logarithmic domain. Experimental results show that the restored images by the proposed method have ISNR improvement of 3.3-4.0 ㏈ and ${\beta}$, a measurement parameter on edge preservation, improvement of 0.0103-0.0126 and superior subjective image quality over those by conventional methods.

• Journal of the Institute of Electronics Engineers of Korea SP
• /
• v.48 no.1
• /
• pp.33-45
• /
• 2011

In this paper, we present a method for extracting video objects efficiently by using the modified graph cut algorithm based on contour information. First, we extract objects at the first frame by an automatic object extraction algorithm or the user interaction. To estimate the objects' contours at the current frame, motion information of objects' contour in the previous frame is analyzed. Block-based histogram back-projection is conducted along the estimated contour point. Each color model of objects and background can be generated from back-projection images. The probabilities of links between neighboring pixels are decided by the logarithmic based distance transform map obtained from the estimated contour image. Energy of the graph is defined by predefined color models and logarithmic distance transform map. Finally, the object is extracted by minimizing the energy. Experimental results of various test images show that our algorithm works more accurately than other methods.

• KSCI Review
• /
• v.2 no.1
• /
• pp.55-67
• /
• 1995

Recently, it is possible to high speed transmission of the non-voiced data, including voice, data, moving image instead of voice only in the past by changing the communication method to digital form from analog owing to the development of semiconductor and computer technology which for information transmission of the remote point. By doing so, we can get the improvement of the noise effect and low cost but the loss of transmission bandwidth. It is necessary to take some method in oreder to reducing the fading which is propotional to transmission bandwidth during the transmission of radio communication channel, especially. When we design the digital communication system, we must considered to the fading effect in order to determination of the transmitting power, modulation /demodulation method, transmission speed, bit error rate. This paper mainly concerns to the method to the channel simulator which descrives the fading effect during the transmission by computer model and digital filter modeling of the radio fading channel by unsing the transmitting and received signal. By taking the inverse of the characteristic of the modeled filter, it is possible to improvement of the communication system by reducing the distortion and inter-symbol interference which occurs in the channel.

• The Journal of the Acoustical Society of Korea
• /
• v.38 no.4
• /
• pp.433-441
• /
• 2019

In this paper, a method of estimating speech parameters for ultrasonic Doppler signals reflected from the articulatory muscles using LSTM (Long Short Term Memory) RNN (Recurrent Neural Networks) was introduced and compared with the method using MLP (Multi-Layer Perceptrons). LSTM RNN were used to estimate the Fourier transform coefficients of speech signals from the ultrasonic Doppler signals. The log energy value of the Mel frequency band and the Fourier transform coefficients, which were extracted respectively from the ultrasonic Doppler signal and the speech signal, were used as the input and reference for training LSTM RNN. The performance of LSTM RNN and MLP was evaluated and compared by experiments using test data, and the RMSE (Root Mean Squared Error) was used as a measure. The RMSE of each experiment was 0.5810 and 0.7380, respectively. The difference was about 0.1570, so that it confirmed that the performance of the method using the LSTM RNN was better.