• Journal of Korea Multimedia Society
• /
• v.6 no.2
• /
• pp.247-258
• /
• 2003

Principal component analysis (PCA) based algorithms form the basis of numerous algorithms and studies in the face recognition literature. PCA is a statistical technique and its incorporation into a face recognition system requires numerous design decisions. We explicitly take the design decisions by in-troducing a generic modular PCA-algorithm since some of these decision ate not documented in the literature We experiment with different implementations of each module, and evaluate the different im-plementations using the September 1996 FERET evaluation protocol (the do facto standard method for evaluating face recognition algorithms). We experiment with (1) changing the illumination normalization procedure; (2) studying effects on algorithm performance of compressing images using JPEG and wavelet compression algorithms; (3) varying the number of eigenvectors in the representation; and (4) changing the similarity measure in classification process. We perform two experiments. In the first experiment, we report performance results on the standard September 1996 FERET large gallery image sets. The result shows that empirical analysis of preprocessing, feature extraction, and matching performance is extremely important in order to produce optimized performance. In the second experiment, we examine variations in algorithm performance based on 100 randomly generated image sets (galleries) of the same size. The result shows that a reasonable threshold for measuring significant difference in performance for the classifiers is 0.10.

• Journal of the Korean Society of Radiology
• /
• v.13 no.4
• /
• pp.623-628
• /
• 2019

The aim of this study was to design fast non local means (FNLM) noise reduction algorithm and to confirm its application feasibility in light microscopic image. For that aim, we acquired mouse first molar image and compared between previous widely used noise reduction algorithm and our proposed FNLM algorithm in acquired light microscopic image. Contrast to noise ratio, coefficient of variation, and no reference-based evaluation parameter such as natural image quality evaluator (NIQE) and blind/referenceless image spatial quality evaluator (BRISQUE) were used in this study. According to the result, our proposed FNLM noise reduction algorithm can achieve excellent result in all evaluation parameters. In particular, it was confirmed that the NIQE and BRISQUE evaluation parameters for analyzing the overall morphologcal image of the tooth were 1.14 and 1.12 times better than the original image, respectively. In conclusion, we demonstrated the usefulness and feasibility of FNLM noise reduction algorithm in light microscopic image of small animal tooth.

• Journal of Internet of Things and Convergence
• /
• v.8 no.4
• /
• pp.41-47
• /
• 2022

Smoothing is a transmission plan that converts video data stored at a variable bit rate into a fixed bit rate. Algorithms for smoothing include CBA, which aims to minimize the number of transmission rate increases, MCBA, which minimizes the number of transmission rate changes, and MVBA algorithms that minimize the amount of transmission rate change. This paper compares the proposed algorithm with the CBA algorithm with various video data, buffer size, and performance evaluation factors as a follow-up to the proposed smoothing algorithm that minimizes (maximizes) the transmission rate increase (decrease) when the server requires more bandwidth The evaluation factors used were compared with the number of changes in the fps rate, the minimum fps, the average fps, fps variability, and the number of frames to be discarded. As a result of the comparison, the proposed algorithm showed superiority in comparing the number of fps rate changes and the number of frames discarded.

• Progress in Medical Physics
• /
• v.23 no.4
• /
• pp.252-260
• /
• 2012

The reduction of radiation dose from x-ray is a main concern in computed tomography (CT) imaging due to the side-effect of the dose on human body. Recently, the various methods for dose reduction have been studied in CT and one of the method is a iterative reconstruction based on total variation (TV) minimization at few-views data. In this paper, we evaluated the image quality between total variation (TV) minimization algorithm and Feldkam-Davis-kress (FDK) algorithm in micro computed tomography (CT). To evaluate the effect of TV minimization algorithm, we produced a cylindrical phantom including contrast media, water, air inserts. We can acquire maximum 400 projection views per rotation of the x-ray tube and detector. 20, 50, 90, 180 projection data were chosen for evaluating the level of image restoration by TV minimization. The phantom and mouse image reconstructed with FDK algorithm at 400 projection data used as a reference image for comparing with TV minimization and FDK algorithm at few-views. Contrast-to-noise ratio (CNR), Universal quality index (UQI) were used as a image evaluation metric. When projection data are not insufficient, our results show that the image quality of reconstructed with TV minimization is similar to reconstructed image with FDK at 400 view. In the cylindrical phantom study, the CNR of TV image was 5.86, FDK image was 5.65 and FDK-reference was 5.98 at 90-views. The CNR of TV image 0.21 higher than FDK image CNR at 90-views. UQI of TV image was 0.99 and FDK image was 0.81 at 90-views. where, the number of projection is 90, the UQI of TV image 0.18 higher than FDK image at 90-views. In the mouse study UQI of TV image was 0.91, FDK was 0.83 at 90-views. the UQI of TV image 0.08 higher than FDK image at 90-views. In cylindrical phantom image and mouse image study, TV minimization algorithm shows the best performance in artifact reduction and preserving edges at few view data. Therefore, TV minimization can potentially be expected to reduce patient dose in clinics.

• Proceedings of the Korea Multimedia Society Conference
• /
• 2001.11a
• /
• pp.681-685
• /
• 2001

본 논문에서는 난수 발생기로써 랜덤성이 우수하고 하드웨어 설계시 고속성이 보장되어 LFSR(Linear Feedback Shift register) 대안으로 소개되고 있는 셀룰라 오토마타(CA: Cellular Automata)를 소개하고 이를 이용한 새로운 블록 암호 알고리즘을 제안한다. 제안된 블록 암호 알고리즘은 Fiestel 구조로써 라운드 함수와 키 스케쥴링 모두를 셀룰러 오토마타를 이용하여 구성함으로써 구현의 편이성과 고속성을 추구하였다. 제안 알고리즘에 대한 간단한 통계적 검정과 성능평가를 통해 기존 표준 알고리즘고의 비교를 수행하였다.

• Proceedings of the Korea Institutes of Information Security and Cryptology Conference
• /
• 2001.11a
• /
• pp.20-25
• /
• 2001

블록 암호 알고리즘과 블록 암호의 운영모드 및 메시지 인증 알고리즘에 대한 의사 난수성의 증명은 근래의 암호 분석에서 구조적인 안전성을 증명을 위한 기법으로 커다란 안전성의 평가 방법으로 자리 매김하고 있다. 본 논문에서는 비동기식(W-CDMA) 3세대 이동통신 3GPP에서의 MAC 알고리즘의 안전성을 의사 난수성에 기반한 concrete security 관점에서의 안전성에 대해 알아본다. 즉, 3GPP-MAC 알고리즘의 구조적인 안전성에 대한 증명을 다룬다.

• Proceedings of the KAIS Fall Conference
• /
• 2009.05a
• /
• pp.520-524
• /
• 2009

본 논문에서는 격자형태의 무선 Ad-hoc 통신망에서 사용하는 포화경로 선정 알고리즘의 성능을 분석하였다. 대체경로 라우팅을 함으로써 좀 더 낮은 차단확률을 기대할 수 있어, 포화경로 선정 알고리즘을 사용의 타당성을 확인할 수 있었다. 또한 기존의 Flood search 알고리즘, Restricted flooding, Hybrid routing의 성능을 모의시험을 통하여 분석한 결과 통신망 효율면에서 Flood search 알고리즘이 우수한 것으로 나타났다.

• 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.

• The Journal of the Acoustical Society of Korea
• /
• v.12 no.1E
• /
• pp.71-77
• /
• 1993

패턴매칭 기법에 근거한 음성 인식 시스템은 크게 clustering 과정과 labeling 과정으로 구성된다. 본 논문에서는 Kohonen의 featrue map 알고리즘과 LVQ2 알고리즘을 각각 clusterer와 labeler로 하는 음소인식 시스템을 구성한다. 구성된 인식시스템의 성능을 향상시키기 위해서 수정된 LVQ2알고리즘(MLVQ2)을 제안한다. MLVQ2는 selective learning, LVQ2, perturbed LVQ2 그리고 기존의 LVQ2의 4단계 학습과정으로 구성된다. 제안된 음소 인식 알고리즘의 성능을 평가하기 위하여 LVQ2와 MLVQ2를 각각 사용하여 6가지의 한국어 음소군에 대한 feature map을 만든다. 음소인식 실험결과, LVQ2와 MLVQ2를 사용하는 경우 각각 60.5%와 65.4%의 인식률을 얻을 수 있었다.

• Proceedings of the Korean Information Science Society Conference
• /
• 2002.10e
• /
• pp.484-486
• /
• 2002

주문형 비디오 서비스(VoD: Video on Demand)의 경우 네트워크의 자원 이용도와QoS (Quality of Service)가 보장되도록 구축되어야 한다. 대역폭 스무딩 알고리즘은 주문형 비디오 시스템에서 비디오 데이터를 보낼 경우 비디오의 데이터의 급격한 대역폭의 증가 또는 감소를 최소한 줄이는 기법이다. 본 논문에서는 기존의 대역폭 스무딩 기법과 비디오 압축 기법의 특성을 이용하여 새로운 모델과 스무딩 알고리즘을 제안한다. 제안된 알고리즘의 성능평가를 위해 기존 스무딩 알고리즘의 결과와 비교 분석한다.