• Proceedings of the KSPS conference
• /
• 2003.05a
• /
• pp.207-210
• /
• 2003

In this paper, the combinations of speech enhancement techniques are experimented. Specifically, the spectral subtraction, KLT based comb-filtering, and their combinations are applied to the Aurora2 database. The results show that recognition accuracy is improved when KLT based comb-filtering is applied after spectral subtraction.

• International Journal of Control, Automation, and Systems
• /
• v.6 no.3
• /
• pp.378-385
• /
• 2008

This paper aims to present a polynomial approach to the steady-state optimal filtering for delayed systems. The design of the steady-state filter involves solving one polynomial equation and one spectral factorization. The key problem in this paper is the derivation of spectral factorization for systems with delayed measurement, which is more difficult than the standard systems without delays. To get the spectral factorization, we apply the reorganized innovation approach. The calculation of spectral factorization comes down to two Riccati equations with the same dimension as the original systems.

• Journal of the Optical Society of Korea
• /
• v.7 no.4
• /
• pp.216-223
• /
• 2003

We report a technique that improves the spatial resolution of optical coherence tomography (OCT) by utilizing fiber-based spectral filtering. The proposed technique improves the resolution by filtering out the erbium’s characteristic peak from the amplified spontaneous emission (ASE) source spectrum, and reshaping the spectrum to Gaussian-like. We used a long period fiber grating (LPG) and an erbium doped fiber (EDF) absorber for the spectral filtering. An in-house made ASE source as well as a commercial ASE source [ASE-FL7002] was used as the OCT sources to study the proposed technique. The resolution of the OCT based on an in-house made ASE source is enhanced from 200 to 40 ㎛ with an LPG. While, the resolution of the OCT based on a commercial ASE source is enhanced from 25 to 19 ㎛ with the aid of an EDF absorber. However, sidelobes still exist in the interferogram due to imperfect spectral filtering, which limited the resolution. Further enhancement in the spatial resolution of the OCT system using the ASE source is possible with the aid of cascaded LPGs and/or carefully designed EDF absorber.

• ETRI Journal
• /
• v.26 no.3
• /
• pp.273-276
• /
• 2004

We propose a novel feature processing technique which can provide a cepstral liftering effect in the log-spectral domain. Cepstral liftering aims at the equalization of variance of cepstral coefficients for the distance-based speech recognizer, and as a result, provides the robustness for additive noise and speaker variability. However, in the popular hidden Markov model based framework, cepstral liftering has no effect in recognition performance. We derive a filtering method in log-spectral domain corresponding to the cepstral liftering. The proposed method performs a high-pass filtering based on the decorrelation of filter-bank energies. We show that in noisy speech recognition, the proposed method reduces the error rate by 52.7% to conventional feature.

• Proceedings of the IEEK Conference
• /
• 2008.06a
• /
• pp.415-416
• /
• 2008

We demonstrate a simple method to multiply the repetition rate of an optical pulse train by a fiber Fabry-Parot interferometer (FFPI) spectral filtering. A stable 40 GHz pulse train at 1550 nm is successfully generated by removing unwanted spectral components of a 10 GHz actively mode-locked laser source by passing a high finesse FFPI.

• Journal of IKEEE
• /
• v.12 no.4
• /
• pp.234-238
• /
• 2008

We demonstrate a simple method to generate a stable 40 GHz pulse train at 1550 nm by spectral filtering of a 10 GHz mode.locked pulse source using a fiber Fabry-Perot interferometer (FFPI). A high finesse FFPI with a 40 GHz free spectral range blocks successfully unwanted spectral components of a 10 GHz pulse source and passes only 40 GHz spaced spectral lines ensuring pulse repetition-frequency quadruplication of the input pulses.

• Proceedings of the Korean Institute of Information and Commucation Sciences Conference
• /
• 2015.05a
• /
• pp.533-536
• /
• 2015

Mobile music market is growing. However, services what are applied recently are inaccurate to recommend music that a user is worth to prefer. So, this paper suggests music recommend system. This system recommend music that users prefer analyzing music information and user's musical propensity and using collaborative filtering. This system classify genre and extract factors what can be get using STFT's ZCR, Spectral roll-off, Spectral flux. So similar musics are clustered by these factors. And then, after divide mood of music's lyric, it finally recommend music automatically using collaborative filtering.

• Proceedings of the Optical Society of Korea Conference
• /
• 2003.07a
• /
• pp.254-255
• /
• 2003

OCT (Optical Coherence Tomography)는 여러 단층영상 촬영기법의 하나로써 절개 없이도 생체의 복잡한 내부구조를 고분해능으로 이미징할 수 있는 장점을 갖고 있다. 고분해능 OCT 이미지는 사용되는 광원에 의해서 결정되는데 가우시안 형태를 갖는 넓은 파장대역의 광원일수록 높은 분해능이 가능하게 된다. OCT에서 사용되는 광원의 종류는 SLD, EELED, ASE등이 사용되고 있다. 그 중에서 ASE광원은 높은 출력 파워를 갖고 넓은 파장대역을 포함하지만 스펙트럼의 형태가 가우시안이 아닌 단점을 가지고 있다. (중략)

• Journal of the Optical Society of Korea
• /
• v.4 no.2
• /
• pp.94-99
• /
• 2000

Optical monitoring using the chromatic aberration of focusing optics is applied to auto-focus control in laser welding. The spectral transmittance of thermal radiation from a weld pool through an aperture depends on the wavelength of the spectral band and on the distance of the weld pool from the focusing optics. Its dependence has been used to monitor the focus shift in laser welding by measuring the spectral band signals filtered by the aperture. The difference between pulsed and continuous laser welding is analyzed. Furthermore, the dependence of the focus shift monitoring on the weld pool size variation is optimized to monitor the focus shift independently from the laser power change at the weld pool. The performance of the auto-focus control with chromatic filtering is presented for pulsed laser welding.

• Korean Journal of Remote Sensing
• /
• v.2 no.2
• /
• pp.79-94
• /
• 1986

TM data, which have better resolution in spatial and spectral than MSS data, were used for this study, and several Image Processing Techniques (IPT) were examined for finding the best IPT to fit to lineament extraction and mineralized zone mapping. The Ryeongnam area was selected as test area, because the area is one of major mineralized zones in Korea and its hydrothermal alteration zone is wider and deeper than other areas. The spatial filtering method is most optimum one for limeament extraction: that is, the directional spatial filtering is most efficient to detect N-S, E-W direction lineaments on the image, and the high boost filtering can be applied for mapping all direction lineaments. The ratio method was selected for detecting altered zone. It is possible to make several tens combinations in ratio with 7 bands of TM data, but considering spectral characteristics of each band of TM to the geological meterials and vegetation, the band 4/band 3(A), band 5/band 7(B), and B/A ratio methods were chosen among them. The 5/7 ratio image did not show clearly the altered area due to noise from vegetation cover, so the 4/3 ratio imae was used for trying to decrease the effect of vegetation. As a result the B/A ratio image showed quite nicely the altered zone of the test area. In conclusion, the spatial filtering is the best image processing techniques for lineament mapping, and the B/A ratio image in TM data is useful for the mineralized zone mapping.