• Steel and Composite Structures
• /
• 제51권6호
• /
• pp.713-727
• /
• 2024

Damage detection in structures using the change of modal parameters (modal shapes and natural frequencies) has achieved satisfactory results. However, as modal shapes and natural frequencies alone may not provide enough information to accurately detect damages. Therefore, a hybrid singular value decomposition and deep belief network approach is developed to effectively identify damages in aluminum plate structures. Firstly, damage locations are determined using singular value decomposition (SVD) to reveal the singularities of measured displacement modal shapes. Secondly, using experimental modal analysis (EMA) to measure the natural frequencies of damaged aluminum plates as inputs, deep belief network (DBN) is employed to search damage severities from the damage evaluation database, which are calculated using finite element method (FEM). Both simulations and experimental investigations are performed to evaluate the performance of the presented hybrid method. Several damage cases in a simply supported aluminum plate show that the presented method is effective to identify multiple damages in aluminum plates with reasonable precision.

• 한국음향학회지
• /
• 제21권5호
• /
• pp.494-502
• /
• 2002

본 연구는 실측 잔향음 자료에서 나타나는 단주기적 시변동성 신호 간섭 (interference)을 억제하기 위해 Ecart-Young 이론을 토대로 자료 행렬로부터 낮은 계수를 추출하여 근사화하는 낮은 계수 근사법 (LRA: Low Rank Approximation) 기법을 제안하였다. 이 기법을 실측 자료에 적용한 결과, 잔향음 신호와 시변동성 신호가 분리되었으며 이때 적절한 낮은 계수를 추출키 위해서 특이치 분해법 (SVD: Singular Value Decomposition)이 사용되었다. 잔향음 신호의 억제는 LRA를 통해 얻어진 근사치와 실측치 사이의 잔차를 계산함으로써 수행하였으며 결과적으로 LRA을 이용하여 시간적으로 안정적인 잔향음 신호를 획득함으로써 능동 소오나 시스템 운용 및 잔향음 모델링시 적용 가능성을 제시하였다.

• Bulletin of the Korean Chemical Society
• /
• 제24권7호
• /
• pp.967-970
• /
• 2003

A new modified singular value decomposition method, piecewise polynomial truncated SVD (PPTSVD), which was originally developed to identify discontinuity of the earth's radial density function, has been used for large solvent peak suppression and noise elimination in nuclear magnetic resonance (NMR) signal processing. PPTSVD consists of two algorithms of truncated SVD (TSVD) and L₁ problems. In TSVD, some unwanted large solvent peaks and noise are suppressed with a certain soft threshold value, whereas signal and noise in raw data are resolved and eliminated in L₁ problems. These two algorithms were systematically programmed to produce high quality of NMR spectra, including a better solvent peak suppression with good spectral line shapes and better noise suppression with a higher signal to noise ratio value up to 27% spectral enhancement, which is applicable to multidimensional NMR data processing.

• 한국정보과학회논문지:소프트웨어및응용
• /
• 제29권4호
• /
• pp.248-257
• /
• 2002

군집화는 객체들의 특성을 분석하여 유사한 성질을 갖고 있는 객체들을 동일한 집단으로 분류하는 방법이다. 전자 상거래 자료처럼 차원 수가 많고 누락 값이 많은 자료의 경우 입력 자료의 차원축약, 잡음제거를 목적으로 SVD를 사용하여 군집화를 수행하는 것이 효과적이지만, SVD를 통해 변환된 자료는 원래의 속성 정보를 상실하기 때문에 군집 결과분석에서 원본 속성의 가치 해석이 어렵다. 따라서 본 연구는 군집화 수행 후 엔트로피 가중치 및 SVD를 이용하여 군집의 중요한 속성을 발견하기 위한 군집 특징 선택 기법 ENTROPY-SVD를 제안한다. ENTROPY-SVD는 자료의 속성들과 유사객체 군과의 묵시적인 은닉 구조를 활용하기 위하여 SVD를 이용하고 유사객체 군에 포함된 응집도가 높은 속성들을 발견하기 위하여 엔트로피 가중치를 사용한다. 또한 ENTROPY-SVD를 적용한 모델 기반의 협력적 여과기법의 추천 시스템 CFS-CF를 제안하고 그 효용성 및 효과를 평가한다.

• 전자공학회논문지S
• /
• 제35S권3호
• /
• pp.176-184
• /
• 1998

We present a sequential factorization method using singular value decomposition (SVD) for recovering both the three-dimensional shape of an object and the motion of camera from a sequence of images. We employ paraperpective projection [6] for camera model to handle significant translational motion toward the camera or across the image. The proposed mthod not only quickly gives robust and accurate results, but also provides results at each frame becauseit is a sequential method. These properties make our method practically applicable to real time applications. Considerable research has been devoted to the problem of recovering motion and shape of object from image [2] [3] [4] [5] [6] [7] [8] [9]. Among many different approaches, we adopt a factorization method using SVD because of its robustness and computational efficiency. The factorization method based on batch-type computation, originally proposed by Tomasi and Kanade [1] proposed the feature trajectory information using singular value decomposition (SVD). Morita and Kanade [10] have extenened [1] to asequential type solution. However, Both methods used an orthographic projection and they cannot be applied to image sequences containing significant translational motion toward the camera or across the image. Poleman and Kanade [11] have developed a batch-type factorization method using paraperspective camera model is a sueful technique, the method cannot be employed for real-time applications because it is based on batch-type computation. This work presents a sequential factorization methodusing SVD for paraperspective projection. Initial experimental results show that the performance of our method is almost equivalent to that of [11] although it is sequential.

• ETRI Journal
• /
• 제33권4호
• /
• pp.652-655
• /
• 2011

In this letter, we propose a new compression method for a high dimensional support vector machine (SVM). We used singular value decomposition (SVD) to compress the norm part of a radial basis function SVM. By deleting the least significant vectors that are extracted from the decomposition, we can compress each vector with minimized energy loss. We select the compressed vector dimension according to the predefined threshold which can limit the energy loss to design criteria. We verified the proposed vector compressed SVM (VCSVM) for conventional datasets. Experimental results show that VCSVM can reduce computational complexity and memory by more than 40% without reduction in accuracy when classifying a 20,958 dimension dataset.

• 한국멀티미디어학회논문지
• /
• 제12권6호
• /
• pp.785-793
• /
• 2009

This paper proposes a signature using dominant singular values for video sequence matching. By considering the input image as matrix A, a partition procedure is first performed to separate the matrix into non-overlapping sub-images of a fixed size. The SVD(Singular Value Decomposition) process decomposes matrix A into a singular value-singular vector factorization. As a result, singular values are obtained for each sub-image, then k dominant singular values which are sufficient to discriminate between different images and are robust to image size variation, are chosen and normalized as the signature for each block in an image frame for matching between the reference video clip and the query one. Experimental results show that the proposed video signature has a better performance than ordinal signature in ROC curve.

• 대한전자공학회:학술대회논문집
• /
• 대한전자공학회 2001년도 하계종합학술대회 논문집(3)
• /
• pp.47-50
• /
• 2001

This paper took advantage of SVD (Singular value Decomposition) techniques of LSI(Latent Semantic Indexing) to grasp easily terminology distribution. Existent LSI did to static database, propose that apply to dynamic database in this paper. But, if dynamic applies LSI to database, updating problem happens. Existent updating way is Recomputing method, Folding-in method, SVD-updating method. Proposed QR decomposition method to show performance improvement than existent three methods in this paper.

• 한국산학기술학회논문지
• /
• 제14권5호
• /
• pp.2393-2399
• /
• 2013

본 논문에서는 특이값 분해(SVD: singular value decomposition)에 의한 기저 스펙트럼의 합성을 통해 혈소판 라만 스펙트럼에서 알츠하이머병(AD: Alzheimer's disease)을 검출하는 방법을 제안하였다. AD가 유도된 형질 전환 실험용 쥐의 혈소판에서 측정한 라만 스펙트럼은 가산 잡음과 배경 잡음의 제거와 정규화로 구성된 전처리 과정을 수행한다. 각 데이터 행렬의 열벡터는 AD와 정상(NR: normal)의 라만 스펙트럼으로 구성한다. 이 데이터 행렬을 SVD로 분해한 다음 각 행렬의 열벡터 12개를 AD와 NR의 기저 스펙트럼으로 결정한다. 분류 과정은 각 클래스의 기저 스펙트럼을 선형 합성한 스펙트럼과 분류 스펙트럼의 평균제곱근오차(root mean square error)가 최소인 클래스를 선택하는 것으로 완료된다. 278개의 혈소판 라만 스펙트럼을 사용한 실험에 따르면 제안한 방법의 평균 분류율은 약 97.6%로 주성분 분석(principle components analysis)으로 추출한 특징에 MLP(multi-layer perceptron)를 이용한 경우보다 약 6.1% 정도의 우수한 성능을 보였다. 이 결과에서 SVD에 의한 기저 스펙트럼이 혈소판 라만 스펙트럼에서 AD의 검출에 적합하게 사용될 수 있음을 확인하였다.

• 대한방사선기술학회지:방사선기술과학
• /
• 제42권2호
• /
• pp.119-130
• /
• 2019

Positron emission tomography (PET) is widely used imaging modality for effective and accurate functional testing and medical diagnosis using radioactive isotopes. However, PET has difficulties in acquiring images with high image quality due to constraints such as the amount of radioactive isotopes injected into the patient, the detection time, the characteristics of the detector, and the patient's motion. In order to overcome this problem, we have succeeded to improve the image quality by using the dynamic image reconstruction method based on singular value decomposition. However, there is still some question about the characteristics of the proposed technique. In this study, the characteristics of reconstruction method based on singular value decomposition was estimated over computational simulation. As a result, we confirmed that the singular value decomposition based reconstruction technique distinguishes the images well when the signal - to - noise ratio of the input image is more than 20 decibels and the feature vector angle is more than 60 degrees. In addition, the proposed methode to estimate the characteristics of reconstruction technique can be applied to other spatio-temporal feature based dynamic image reconstruction techniques. The deduced conclusion of this study can be useful guideline to apply medical image into SVD based dynamic image reconstruction technique to improve the accuracy of medical diagnosis.