• Proceedings of the Korea Information Processing Society Conference
• /
• 2001.10a
• /
• pp.605-608
• /
• 2001

다층 신경망의 학습에 쓰이는 오류 역전파 학습은 매우 효과적이지만 학습 속도가 너무 느리고 최적의 은닉충의 뉴런의 수를 결정하는 해답은 아직 없는 실정이다. 또한 가끔은 국부 최소점(Local maxima)에 빠져 학습이 끝내 이루어지지 않는 경우가 있다. 이에 본 논문에서는 이러한 Local maxima 를 효과적으로 탈출 할 수 있는 방법에 대해서 연구해 보았다. 국부 최소점은 연결강도와 전체 오차 사이의 이차원 공간에서 표현할 수 있는데 본 알고리즘은 이러한 연결강도와 오차와의 관계를 인위적으로 변화시켜 결론적으로 Local maxima 를 탈출하게 하는 방법을 소개한다. 본 연구에서 사용된 방법은 네트웍이 학습중에 Local maxima 에 빠졌을 때 은닉층의 뉴런의 수를 추가하여 인위적으로 연결강도 평면의 위상을 변조시킨다. 또한 은닉충의 뉴런의 수를 동적으로 변화 시키면서 최적의 뉴런의 수를 결정할 수 있게 하였다. 위 알고리즘의 성능을 평가하기 위해서 XOR 문제와 $10{\times}8$ 영문폰트와 숫자의 학습에 적용하여 일반적인 역전파 학습과 비교 평가하였다.

• Proceedings of the Korean Information Science Society Conference
• /
• 2001.10b
• /
• pp.313-315
• /
• 2001

본 논문에서 소개하는 알고리즘을 은닉층의 뉴런의 수를 학습하는 동안 동적으로 변화시켜 역전파 알고리즘의 단점인 Local maxima를 탈출하고 또한 은닉층의 뉴런의 수를 결정하는 과정을 없애기 위해 연구되었다. 본 알고리즘의 성능을 평가하기 위해 두 가지 실험에 적용하였는데 첫번째는 Exclusive-OR 문제이고 두번째는 7$\times$8 한글 자음과 모음의 폰트 학습에 적용하였다. 이 실험의 결과로 네트웍이 local maxima에 빠져드는 확률이 줄어드는 것을 알 수 있었고 학습속도 또한 일반적인 역전파 알고리즘보다 빠른 것으로 증명되었다.

• Journal of the Institute of Electronics Engineers of Korea SP
• /
• v.41 no.6
• /
• pp.109-120
• /
• 2004

The current paper presents an effective deblocking algorithm for block-based coded images using singularity detection in Mallat wavelet transform. In block-based coded images. The local maxima of the wavelet transform modulus detect all singularities, including blocking artifacts, from multiscale edges. Accordingly, the current study discriminates between blocking artifacts and edges by estimating the Lipschitz regularity of the local maxima and removing the wavelet transform modulus of blocking artifacts. Experimental results showed that the performance of the proposed algorithm was objectively and subjectively superior.

• Journal of the Institute of Electronics Engineers of Korea SP
• /
• v.48 no.3
• /
• pp.99-105
• /
• 2011

In this paper, we proposed a method for estimating an accurate gamut based on segment maxima method. According to the number of segments in the segment maxima, a local concavity is generated in the vicinity of lightness axis or a gamut is reduced in high chroma region. It induces artifacts or deterioration of the image quality. To remove these artifacts, the number of segment is determined according to the number of samples. and a local concavity is modified by extending a detected concave point to the line connecting two adjacent boundary points. Experimental results show that the contours in a uniform color region and speckle artifacts from the conventional segment maxima algorithm are removed.

• Journal of Korean Society for Geospatial Information Science
• /
• v.14 no.4 s.38
• /
• pp.19-26
• /
• 2006

This study assessed the accuracies of the ground information extracting methods from the LiDAR data. Especially, it compared two kinds of method, one of them is using directly the raw LiDAR data which is point type vector data and the other is using changed data to DSM type as the normal grid type. The methods using Local Maxima and Entropy methods are applied as a former case, and for the other case, this study applies the method using edge detection with filtering and the generated reference surface by the mean filtering. Then, the accuracy assessment are performed with these results, DEM constructed manually and the error permitted limit in scale of digital map. As a results, each DEM mean errors of methods using edge detection with filtering, reference surface, Local Maxima and Entropy are 0.27m, 2.43m, 0.13m and 0.10m respectively. Hence, the method using entropy presented the highest accuracy. And an accuracy from a method directly using the raw LiDAR data has higher accuracy than the method using changed data to DSM type relatively.

• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.27 no.10A
• /
• pp.1011-1020
• /
• 2002

The current paper presents an effective deblocking algorithm for block-based coded images using singularity detection in a wavelet transform. In block-based coded images, the local maxima of a wavelet transform modulus detect all singularities, including blocking artifacts, from multiscale edges. Accordingly, the current study discriminates between a blocking artifact and an edge by estimation the Lipschitz regularity of the local maxima and removing the wavelet transform modulus of a blocking artifact that has a negative Lipschitz regularity exponent. Experimental results showed that the performance of the proposed algorithm was objectively and subjectively superior.

• Journal of IKEEE
• /
• v.24 no.4
• /
• pp.1081-1085
• /
• 2020

Instantaneous frequencies (IF) of the AM-FM signal is estimated based on the inflection point detection (IPD) method. Local maxima/minima are detected using the IPD, and they are exploited to find the IF of AM and FM components, respectively. The envelope of the maxima/minima is obtained to estimate the IF of the AM part. And the distance between neighboring maxima (or minima) is used to estimate the IF of the FM component. Computer simulation shows that the proposed method properly estimates the IF of the AM and FM when the signal has fixed frequencies for both parts. In the case of the time-varying IF of the FM part, the estimated IF shows some deviation from the true IF due to the rough sampling effect of the maximum/minimum points. Thus, the post-processing such as the lowpass filtering of the estimated IF is required to refine the resulting IF estimation.

• Bulletin of the Korean Chemical Society
• /
• v.18 no.6
• /
• pp.573-578
• /
• 1997

Ion fragmentations of fatty acids such as stearic acid, palmitic acid, myristic acid, and lauric acid were studied using mass spectrometry and semiempirical calculations. The mass spectra of fatty acids showed the distributions of CH₃(CH₂)$_n^+$ and $[(CH_2)_nCO_2H]^+$ fragment ions. The relative ion abundance distributions of $[(CH_2)_nCO_2H]^+$ showed the local maxima at n=6, 10, and 14. The local maximum phenomena were also found in the mass spectra of methyl stearate but not in those of normal alcohols. These local maxima could be explained not by heats of reaction for fragmentation but by the cyclic structures of the molecular ions. The AM1 semiempirical calculations for fatty acids clearly show that the linear structures are more favorable than the cyclic ones for neutral molecules while the cyclic structures are more favorable than the linear ones for ionic molecules. The distances between carboxyl group and methylene of the cylic structures of ionized fatty acids were calculated. The methylene carbons with n=6, 10, and 14 were closer to the carboxyl group than adjacent ones.

• Journal of KIISE:Software and Applications
• /
• v.29 no.3
• /
• pp.202-210
• /
• 2002

During the last few years, the extraction and reconstruction of the blood vessels in the medical image has been actively researched and the analysis for the retinal vessel structure has provided important information for diagnosis and remedy of the retinopathy patients. In this research, we propose the algorithm that tracks automatically the entire retinal vessel in retinal image acquired by the ICG(IndoCyanine Green) technology. This algorithm extracts contours and centers by estimating the local maxima and processing directions and detects bifurcations and junctions by comparing direction components of the local maxima from the gradient magnitude profile of each blood vessel. We present experimental results that the entire blood vessel is automatically reconstructed and is excellent in accuracy and connectivity after applying our algorithm to the ICG retinal images of patients.

• Journal of the Korea Society of Computer and Information
• /
• v.19 no.8
• /
• pp.107-117
• /
• 2014

This paper proposed the method of removing baseline drift by eliminating local maxima such as P, R, T-wave signal region and local minima Q, S-wave signal region. We applied morphology-pair operations improved from morphology operation to the ECG signal. To eliminate overshoot in the result of morphology-pair operation, we apply median value operation to the result of morphology-pair operation. We use MIT/BIH database to estimate the proposed algorithm. Experiment result show that proposed algorithm removing baseline drift effectively without orignal ECG signal distortion.