• Proceedings of the Korea Multimedia Society Conference
• /
• 2004.05a
• /
• pp.546-549
• /
• 2004

본 논문에서는 SVM에 기반 하여 template matching 기법을 도입한 얼굴 검출 알고리즘을 제안한다. 또한 얼굴의 기하학적 특징에 기반 해 얼굴일 확률이 낮은 경우 비 얼굴로 인식하는 방법으로 SVM의 적용 횟수를 줄여 검출속도를 향상 시켰다. 그리고 single response criterion을 적용하여 얼굴을 중복검출하거나 얼굴 영역 내에서 얼굴을 찾는 경우를 제거하여 속도를 개선하였다.

• Proceedings of the Korean Society of Fisheries Technology Conference
• /
• 2003.10a
• /
• pp.35-38
• /
• 2003

현재, 세계 여러나라에서 해상물표를 정확하게 탐지 및 검출하기 위한 방안으로 레이더 clutter 신호를 효과적으로 억제 및 제거하기 위한 연구가 활발하게 진행되고 있다. 일반적으로 레이더 반사파에 대한 envelope 신호의 진폭는 Rayleigh 분포에 따라 변동하는 특성을 나타내지만, clutter의 진폭분포의 파라 메터가 변동하여 분포형상이 변화하면, 오경보확률(false alarm probability)에도 변화가 발생하기 때문에 오경보확률을 충분히 낮은 일정치로써 억제시켜 일정오경보확률(constant false alarm rate, CFAR)을 유지하는 처리가 필요하다. (중략)

• The Journal of the Acoustical Society of Korea
• /
• v.27 no.7
• /
• pp.372-378
• /
• 2008

Voice activity detectors (VAD) are important in wireless communication and speech signal processing, In the conventional VAD methods, an expression for the likelihood ratio test (LRT) based on statistical models is derived in discrete Fourier transform (DFT) domain, Then, speech or noise is decided by comparing the value of the expression with a threshold, This paper presents a new statistical VAD method based on a signal subspace approach, The probabilistic principal component analysis (PPCA) is employed to obtain a signal subspace model that incorporates probabilistic model of noisy signal to the signal subspace method, The proposed approach provides a novel decision rule based on LRT in the signal subspace domain, Experimental results show that the proposed signal subspace model based VAD method outperforms those based on the widely used Gaussian distribution in DFT domain.

• Journal of the Institute of Electronics Engineers of Korea TC
• /
• v.42 no.8 s.338
• /
• pp.33-40
• /
• 2005

This paper presents a structure of the searcher using a diversity in array antenna systems operating in the cdma2000 1x signal environments. The new technique exploits the fact that the In-phase and quadrature components of interferers can respectively be viewed as an independent gaussian noise at each antnna element in most practical cdma signal environments. The proposed PN acquisition scheme is a singles-dwell PN acquisition system consisting of two stages, that is, the searching stage and the verification stage. The searching stage independently correlates the receiver multiple signals with PN generator of each antenna element for obtaining the synchronous energy at the entire region. Then, the searching results of each antenna element are non-coherently combinind. The verification stage compares the searching energy with the optimal threshold, which is predesigned in the lock detector, and decides whether the acquisition is successful or fail. In this paper, we analyzed the effect of tile diversity order to determine the mean acquisition time. In general, it is known that the mean acquisition time significantly decrease as the number of antenna elements increases. But, as the diversity order goes up, the enhancement of the performance is saturated. Therefore, to decrease the mean acquisition time of the searcher, we must design the optimal array antenna systems by considering the operating SNR range of the receiver, the probability of detection $P_D$ and that of false alarm $P_{FA}$ . The Performance of the proposed PN acquisition scheme is analyzed in frequency selective Rayleigh fading channels. In this paper, the effect of the number of antenna elements on PN acquisition scheme is shown according to the probability of detection $P_D$ and that of false alarm $P_{FA}$.

• Journal of the Korea Society of Computer and Information
• /
• v.6 no.3
• /
• pp.78-83
• /
• 2001

In this paper, we are considered Nakagami-m fading, which can model variable multipath mobile radio communication channel, in DS-CDMA system. System modeling using nakagami -m fading is suited for urban mobile communication channel with multipath. We used adaptive serial search PN code acquisition scheme and derived the detection probability($P_D$) and false alarm probability($P_FA$) which have influence on code acquisition time, over Nakagami-m fading. Detection probability($P_D$) and false alarm probability($P_FA$) are detection variable to decide PN code acquisition time and should use to calculate mean and variance. of acquisition time. From computer simulation, we analyzed mean and variance about PN code acquisition of fading channel. Then we can apply it to the H/W design of mobile communication.

• Journal of KIISE
• /
• v.41 no.11
• /
• pp.921-926
• /
• 2014

Face detection refers to the process extracting facial regions in an input image, which can improve speed and accuracy of recognition or authorization system, and has diverse applicability. Since conventional works have tried to detect faces based on the whole shape of faces, its detection performance can be degraded by occlusion made with accessories or parts of body. In this paper we propose a method combining local feature descriptors and probability modeling in order to detect partially occluded face effectively. In training stage, we represent an image as a set of local feature descriptors and estimate a statistical model for normal faces. When the test image is given, we find a region that is most similar to face using our face model constructed in training stage. According to experimental results with benchmark data set, we confirmed the effect of proposed method on detecting partially occluded face.

• The Journal of the Korea institute of electronic communication sciences
• /
• v.10 no.8
• /
• pp.901-906
• /
• 2015

We presented a method for Vad(Voice Activity Detection) using Bi-level HMM. Conventional methods need to do an additional post processing or set rule-based delayed frames. To cope with the problem, we applied to VAD a Bi-level HMM that has an inserted state layer into a typical HMM. And we used posterior ratio of voice states to detect voice period. Considering MFCCs(: Mel-Frequency Cepstral Coefficients) as observation vectors, we performed some experiments with voice data of different SNRs and achieved satisfactory results compared with well-known methods.

• Journal of Advanced Navigation Technology
• /
• v.12 no.2
• /
• pp.132-138
• /
• 2008

Recently, many frequency sharing technology including CR (Cognitive Radio) technology have been studied actively because of frequency scarcity. For implementing CR systems, spectrum sensing technology plays key role. In IEEE 802.22 WRAN, the first standard aimed at Internet service in a TV channel with CR technology. This paper explained the various spectrum sensing schemes detecting narrowband ATSC signal in the IEEE 802.22 such as Energy detector, Spectral Correlation Sensing scheme, Covariance based Sensing algorithm and analyzed in terms of the probability of misdetection when the probability of false alarm is 10%. Based on simulation results, this paper proposed a duplex sensing scheme based on Energy detector and analyzed the capability of the proposed sensing scheme comparing with each sensing scheme.

• The Journal of the Acoustical Society of Korea
• /
• v.30 no.7
• /
• pp.396-401
• /
• 2011

In this paper, we propose a novel voice activity detection (VAD) algorithm to effectively distinguish speech from nonspeech in various noisy environments. The presented VAD utilizes the power spectral deviation (PSD) based on Teager energy (TE) instead of the conventional PSD scheme to improve the performance of decision for speech segments. In addition, the speech absence probability (SAP) is derived in each frequency subband to modify the PSD for further VAD. Performances of the proposed VAD algorithm are evaluated by objective test under various environments and better results compared with the conventional methods are obtained.

• Journal of the Korean Institute of Telematics and Electronics
• /
• v.25 no.10
• /
• pp.1244-1251
• /
• 1988

Edge is one of the primitive features of an image and is widely used in image classification and analysis. New edge extration methods using central moments are presented and show various characteristics according to the order of moment, definition of both random variables and probability density functions. The proposed methods use the integral of differences between local mean and pixels in the window whereas most of conventional edge operators use only differential concepts. This gives good noise immunity and extracts fine edges.