• Journal of the Institute of Electronics Engineers of Korea CI
• /
• v.42 no.6
• /
• pp.83-92
• /
• 2005

This paper presents a watermarking approach that the 1-level Discrete Wavelet Transform(DWT) coefficients of a $64{\ast}64$ binary logo image as watermarks are inserted in LL band and other specific frequency bands of the host image using Multi-Resolution Analysis(MRA) Wavelet transform for copyright protection of image data. The DWT coefficients of the binary logo image are inserted in blocks of LL band and specific bands of the host image that the 3-level DWT has been performed in the same orientation. We investigate Significant Coefficients(SCs) in each block of the frequency areas in order to prevent the quality deterioration of the host image and the watermark is inserted by SCs. When the host image is distorted by difference of the distortion degree in each frequency, we set the thresholds of SCs on each frequency and completely insert the watermark in each frequency of the host image. In order to be invisibility of the watermark, the Human Visual System(HVS) is applied to the watermark. We prove the proper embedding method by experiment. Thereby, we rapidly detect the watermark using this watermarking method and because the small size watermarks are inserted by HVS and SCs, the results confirm the superiority of the proposed method on invisibility and robustness.

• Journal of KIISE:Computing Practices and Letters
• /
• v.8 no.3
• /
• pp.289-294
• /
• 2002

A valid speech-sound block can be classified to provide important information for speech recognition. The classification of the speech-sound block comes from the MRA(multi-resolution analysis) property of the DWT(discrete wavelet transform), which is used to reduce the computational time for the pre-processing of speech recognition. The merging algorithm is proposed to extract valid speech-sounds in terms of position and frequency range. It needs some numerical methods for an adaptive DWT implementation and performs unvoiced/voiced classification and denoising. Since the merging algorithm can decide the processing parameters relating to voices only and is independent of system noises, it is useful for extracting valid speech-sounds. The merging algorithm has an adaptive feature for arbitrary system noises and an excellent denoising SNR(signal-to-nolle ratio).

• Proceedings of the Korean Society for Technology of Plasticity Conference
• /
• 2003.05a
• /
• pp.27-33
• /
• 2003

In closed-die forging process, blocker has been used to fill and distribute metal well in finisher die. Generally, the blocker shape was determined by an expert with many experiences. However, the manual blocker design process takes much time and efforts, so various automatic methods for the blocker design process have been suggested for the last three decades. The method with filtering in FFT (Fast Fourier Transform) for the blocker design provides general solution than other methods. But, due to the properties of FFT in time-frequency domain, this method has some drawbacks such as long calculation time, difficulty of local control and additional boundary process after filtering. In this study, DWT (Discrete Wavelet Transform), which is more flexible and is more wildly used than FFT, is applied to the blocker design. The method with filtering in DWT is very proper to design blocker in both 2-D and 3-D shapes. To verify the efficiency of this method, blockers of some models are designed and the results show that blocker design with DWT is effective fer the blocker designs

• The Transactions of the Korean Institute of Power Electronics
• /
• v.11 no.6
• /
• pp.570-576
• /
• 2006

This paper describes a method for the detection of shorted-turn in the cylindrical synchronous generator rotor windings based on the discrete wavelet transform. Multi-resolution analysis(MRA) based on discrete wavelet transform provides a set of decomposed signals in independent frequency bands, which contain independent dynamic information due to the orthogonality of wavelet function. In the proposed method, shorted-turn detection in rotor windings is based on the decomposition of the rotor currents, where wavelet coefficients of these signals have been extracted. Comparing these extracted coefficients is used for diagnosing the healthy machine from faulty machine. Experimental results are presented for healthy, and machines with 25%, 42%, 67%, 83%, 99% inter-turn short circuits in a rotor slot. Deviation of wavelet coefficients in healthy mode from faulty modes depicts the inverse proportion of shorted-turns. Experimental results show the effectiveness of the proposed method for shorted-turn detection in the cylindrical synchronous generator rotor windings.

• Transactions of Materials Processing
• /
• v.12 no.4
• /
• pp.277-283
• /
• 2003

In a closed-die forging process, blocker has been used to fill and distribute metal well in finisher die. Generally, the blocker shape was determined by an expert with many experiences. However, the manual blocker design process takes much time and efforts, so various automatic methods for the blocker design process have been suggested for the last three decades. The method with filtering in FFT (Fast Fourier Transform) for the blocker design provides general solution than other methods. But. due to the properties of FFT in time-frequency domain, this method has some drawbacks such as long calculation time, difficulty of local control and additional boundary process after filtering. In this study. DWT (Discrete Wavelet Transform), which is more flexible and is more wildly used than FFT, is applied to the blocker design. The method with filtering in DWT is very proper to design blocker in both 2-D and 3-D shapes. To verify the efficiency of this method, blockers of some models are designed and the results show that blocker design with DWT is effective for the blocker designs.

• Proceedings of the KIPE Conference
• /
• 2012.11a
• /
• pp.149-150
• /
• 2012

본 논문에서는 배터리 팩을 구성하는 셀간의 전압편차를 이용한 이산 웨이블릿 변환(DWT;discrete wavelet transform) 기반 SOH(State-of-health) 예측방법을 소개한다. 충방전 전압은 DWT의 다해상도 분석(MRA;multi-resolution analysis)을 이용한 시간-주파수 분석을 통해 고주파 전압 성분(detail;$D_n$)과 저주파 전압 성분(approximation;$A_n$)으로 추가 분해되어 SOH 예측을 위한 추가정보를 제공한다. 각 성분의 통계처리(표준편차)를 통해 노화 이전과 이후의 성분값을 비교한다. 즉 프레시 배터리팩과 노화된 팩의 표준편차 기반 셀간 불균형을 서로 비교하여 SOH 예측이 가능하다.

• Proceedings of the KIPE Conference
• /
• 2013.07a
• /
• pp.449-450
• /
• 2013

본 논문에서는 이산 웨이블릿 변환(DWT;discrete wavelet transform)을 이용한 에너지 저장시스템(ESS;energy storage system)의 고장 검출 방법을 제안한다. ESS에 순간적인 고장 발생시 전압의 급격한 변화가 발생할 수 있으며 이는 다해상도 분석(MRA;multi-resolution analysis)을 이용한 시간-주파수 분석을 통해 분해된 저주파 전압 성분(approximation;$A_n$)과 고주파 전압 성분(detail;$D_n$)중 현저한 성분의 변화가 관찰되는 고주파 전압 성분을 선택한다. 이를 검증하기 위하여 모든 고주파 전압 성분의 절대값을 적용한 뒤 최대값 정보를 추출한다. 이 때, 추출된 각 성분의 최대값과 최대값의 평균을 비교하되 여러 사전실험을 통해 정해진 특정 임계값 대비 큰 값을 나타낼 때 고장이 발생하였음을 판단한다.

• Proceedings of the KIPE Conference
• /
• 2014.07a
• /
• pp.150-151
• /
• 2014

높은 SOC(state-of-charge) 추정알고리즘의 성능을 위해서는 측정된 배터리 단자전압의 정확도가 요구된다. 그렇지만, 예기치 않은 에러로 인해 단자전압에 노이즈 성분이 추가될 경우 SOC 추정성능의 저하를 피할 수 없다. 그러므로, 본 논문에서는 이산 웨이블릿 변환(DWT;discrete wavelet transform)의 다해상도 분석(MRA;multi resolution analysis)의 디노이징(denoising)기법을 적용한 이차전지의 SOC 추정방법을 소개한다. MRA의 시간-주파수 분석을 통해 분해(decomposition)된 저주파 성분(approximation;$A_n$)과 고주파 성분(detail;$D_n$)중 노이즈에 관계된 $D_n$의 고주파 상세 계수(detail coefficient) $d_{j,k}$를 새로이 조정하고 이를 합성(synthesis)하여 디노이징을 마무리 한다. 확장 칼만필터(EKF;extended Kalman filter)의 비교 분석을 통해 제안된 방법의 타당성을 검증한다.

• Proceedings of the KIPE Conference
• /
• 2012.07a
• /
• pp.244-245
• /
• 2012

본 논문에서는 이산 웨이블릿 변환(DWT;discrete wavelet transform)의 다해상도 분석(MRA;multi-resolution analysis)을 통해 분해된 배터리의 저주파 전압 성분(approximation;$A_n$) 기반 SOC(State-of-charge) 추정방법을 소개한다. 급격한 전압 변화의 특성을 나타내는 고주파 전압 성분(detail;$D_n$)이 제거되고 저주파 전압 성분만이 SOC 추정을 위해 사용된다. 이 경우 기존 확장 칼만필터(EKF;extended Kalman filter)에서 SOC 추정에러를 개선하기 위해 사용되었던 노이즈 모델의 생략이 가능하여 알고리즘의 복잡성이 개선된다. 개선된 확장 칼만필터 기반 SOC 추정 결과를 통해 제안된 방법을 검증하였다.

• Journal of KIISE:Software and Applications
• /
• v.31 no.6
• /
• pp.799-807
• /
• 2004

In this paper, we deal with an emotion recognition method using facial images and speech signal. Six basic human emotions including happiness, sadness, anger, surprise, fear and dislike are investigated. Emotion recognition using the facial expression is performed by using a multi-resolution analysis based on the discrete wavelet transform. And then, the feature vectors are extracted from the linear discriminant analysis method. On the other hand, the emotion recognition from speech signal method has a structure of performing the recognition algorithm independently for each wavelet subband and then the final recognition is obtained from a multi-decision making scheme.