• KSII Transactions on Internet and Information Systems (TIIS)
• /
• 제13권3호
• /
• pp.1403-1417
• /
• 2019

With the rapid growth of multimedia data, research on cross-media feature learning has significance in many applications, such as multimedia search and recommendation. Existing methods are sensitive to noise and edge information in multimedia data. In this paper, we propose a semi-supervised method for cross-media feature learning by means of $L_{2,q}$ norm to improve the performance of cross-media retrieval, which is more robust and efficient than the previous ones. In our method, noise and edge information have less effect on the results of cross-media retrieval and the dynamic patch information of multimedia data is employed to increase the accuracy of cross-media retrieval. Our method can reduce the interference of noise and edge information and achieve fast convergence. Extensive experiments on the XMedia dataset illustrate that our method has better performance than the state-of-the-art methods.

• 전자공학회논문지CI
• /
• 제46권4호
• /
• pp.102-110
• /
• 2009

본 논문은 적외선영상(infrared image)에서 배경모델링 기반의 실시간 객체 탐지 기법과 고속 PPC(PowerPC) & FPGA(Field Programmable Gate Array) 기반 개방형 구조의 하드웨어 설계 방법을 제안한다. 개방형 구조는 하드웨어 및 소프트웨어의 이식이 용이하고, 확장, 호환성, 관리 및 유지보수 등이 편리한 장점이 있다. 제안된 배경모델링 방법을 개방형 구조에 탑재하기 위하여 입력영상에서 검색영역 템플릿을 성긴 블록으로 구성하여 탐색영역의 크기를 줄인다. 또한, 이전 프레임과 현재 프레임에서 영상의 흔들림이 발생했을 때 보정하기 위해 전역움직임 보상방법을 적용한다. 배경과 객체를 분리는 픽셀 밝기의 시간 분석을 통해 적응적 값을 적용한다. 분리된 객체주변에 발생하는 클러터 제거 방법은 중앙값 필터를 적용한다. 설계된 임베디드 시스템에서 배경모델링, 객체탐지, 중앙값 필터, 라벨링, 합병 등의 방법은 PPC에서 구현하였다. 실험결과 제안된 임베디드 시스템에서 전역 움직임 보정과 배경예측을 통해 실시간으로 객체가 탐지될 수 있음을 보였다.

• 한국통신학회논문지
• /
• 제31권10C호
• /
• pp.1000-1010
• /
• 2006

본 논문은 DCT(Discrete Cosine Transform) 기반의 움직임 예측 및 보상을 위한 새로운 연산 아키텍처를 제안한다. 기존 방식들의 경우 연산 시간의 단축을 위하여 2차원 DCT 계수의 희소성을 충분히 활용하지 못하고 있다. 본 논문에서는 DCT 영역에서의 효율적인 움직임 예측을 위한 재귀 방정식을 유도하고, 이를 바탕으로 PE로 구성된 WAP를 개발한다. 또한, 재귀 방정식을 이용하여, 움직임 예측된 영상이 저주파 성분부터 고주파 성분까지 다양한 주파수 대역을 갖는 것이 가능함을 보인다. WAP는 아키텍처의 수정 없이 로그형 탐색이나 3단계 탐색과 같은 다양한 움직임 예측 알고리즘들을 수행할 수 있으며, 이러한 특성들은 비디오 부호화와 복호화에 필요한 전력 소모를 줄이기 위하여 이용될 수 있다. 본 논문에서 제안한 WAP 아키텍처는 계산의 복잡도와 연산 시간을 효과적으로 감소시키며, SAD기준을 이용한 DCT 영역에서의 움직임 예측 및 보상 방식은 SAD 또는 SSD 기준을 이용한 공간 영역에서의 움직임 예측 및 보상 방식보다 높은 PSNR과 압축률을 제공함을 보여준다.

• 한국BIM학회 논문집
• /
• 제6권1호
• /
• pp.25-32
• /
• 2016

In BIM environments for infrastructures, civil structures such as road, bridge and tunnel etc., created into 3D objects, integrated with their properties, securing BIM design productivity is very critical during 3D modeling. To solve this issue, configuring BIM libraries so that the users can utilize prefabricating in advance 3D objects that have been applying repeatedly during BIM design is essential so much. Current BIM libraries have made focused on Ready-Made for architectural facilities. However, establishing environment for delivering BIM library for civil facilities is very sparse. Accordingly, this study developed a web-based sharing system for delivering BIM library contents based on standardized drawing for civil area. To do this, we have analyzed core features and operation system of BIM library sharing system in domestic and overseas. Besides, functional requirements that are necessary for developing BIM library sharing system was derived, and through operation scenario configuration, the web-based system was developed according to the detailed mechanism. It is expected that this system can enhance BIM design productivity during library based modeling, and can be utilized as construction supporting tools that can help construction managers to make a design change.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• 제7권11호
• /
• pp.2702-2719
• /
• 2013

A new optimal scheme based on curvelet transform is proposed for retinal image enhancement (RIE) using real-coded quantum genetic algorithm. Curvelet transform has better performance in representing edges than classical wavelet transform for its anisotropy and directional decomposition capabilities. For more precise reconstruction and better visualization, curvelet coefficients in corresponding subbands are modified by using a nonlinear enhancement mapping function. An automatic method is presented for selecting optimal parameter settings of the nonlinear mapping function via quantum genetic search strategy. The performance measures used in this paper provide some quantitative comparison among different RIE methods. The proposed method is tested on the DRIVE and STARE retinal databases and compared with some popular image enhancement methods. The experimental results demonstrate that proposed method can provide superior enhanced retinal image in terms of several image quantitative evaluation indexes.

• Interdisciplinary Bio Central
• /
• 제4권4호
• /
• pp.14.1-14.6
• /
• 2012

Splice site prediction in DNA sequence is a basic search problem for finding exon/intron and intron/exon boundaries. Removing introns and then joining the exons together forms the mRNA sequence. These sequences are the input of the translation process. It is a necessary step in the central dogma of molecular biology. The main task of splice site prediction is to find out the exact GT and AG ended sequences. Then it identifies the true and false GT and AG ended sequences among those candidate sequences. In this paper, we survey research works on splice site prediction based on support vector machine (SVM). The basic difference between these research works is nucleotide encoding technique and SVM kernel selection. Some methods encode the DNA sequence in a sparse way whereas others encode in a probabilistic manner. The encoded sequences serve as input of SVM. The task of SVM is to classify them using its learning model. The accuracy of classification largely depends on the proper kernel selection for sequence data as well as a selection of kernel parameter. We observe each encoding technique and classify them according to their similarity. Then we discuss about kernel and their parameter selection. Our survey paper provides a basic understanding of encoding approaches and proper kernel selection of SVM for splice site prediction.

• 한국자기공명학회논문지
• /
• 제20권4호
• /
• pp.102-108
• /
• 2016

With the rapid increase of data on protein-protein interactions, the need for delineating the 3D structures of huge protein complexes has increased. The protocols for determining nuclear magnetic resonance (NMR) structure can be applied to modeling complex structures coupled with sparse experimental restraints. In this report, I suggest the use of multiple rigid bodies for improving the efficiency of NMR-assisted structure modeling of huge complexes using CYANA. By preparing a region of known structure as a new type of residue that has no torsion angle, one can facilitate the search of the conformational spaces. This method has a distinct advantage over the rigidification of a region with synthetic distance restraints, particularly for the calculation of huge molecules. I have demonstrated the idea with calculations of decaubiquitins that are linked via Lys6, Lys11, Lys27, Lys29, Lys33, Lys48, or Lys63, or head to tail. Here, the ubiquitin region consisting of residues 1-70 was treated as a rigid body with a new residue. The efficiency of the calculation was further demonstrated in Lys48-linked decaubiquitin with ambiguous distance restraints. The approach can be readily extended to either protein-protein complexes or large proteins consisting of several domains.

• 대한전기학회:학술대회논문집
• /
• 대한전기학회 1987년도 전기.전자공학 학술대회 논문집(I)
• /
• pp.311-315
• /
• 1987

In this paper an efficient algorithm to estimate the volume and surface area and the reconstruction algorithm for 3-dimensional graphics are presented. The graph theory is used to estimate the optimal quantitative factors. To improve the computing efficiency, the algorithm to get proper contour points is performed by applying several tolerances. The search and the given arc cost is limited according to the change of curvature of the cross-sectional contour. For mathematical model, these algorithms for volume estimation based on polyhedral approximation are applied to the selected optimal surface. The results show that the values of the volume and surface area for tolerances 1.0005, 1.001 and 1.002 approximate to values for tolerances 1.000 resulting in small errors. The reconstructed three-dimensional images are sparse and consist of larger triangular tiles between two cross sections as tolerance is increasing.

• ETRI Journal
• /
• 제46권3호
• /
• pp.392-403
• /
• 2024

Even though generalized frequency division multiplexing is an alternative waveform method expected to replace the orthogonal frequency division multiplexing in the future, its implementation must alleviate channel effects. Least-squares (LS), a low-complexity channel estimation technique, could be improved by using the discrete Fourier transform (DFT) without increasing complexity. Unlike the usage of the LS method, the DFT-based method requires the receiver to know the channel impulse response (CIR) length, which is unknown. This study introduces a simple, yet effective, CIR length estimator by utilizing LS estimation. As the cyclic prefix (CP) length is commonly set to be longer than the CIR length, it is possible to search through the first samples if CP is larger than a threshold set using the remaining samples. An adaptive scale is also designed to lower the error probability of the estimation, and a simple signal-to-interference-noise ratio estimation is also proposed by utilizing a sparse preamble to support the use of the scale. A software simulation is used to show the ability of the proposed system to estimate the CIR length. Due to shorter CIR length of rural area, the performance is slightly poorer compared to urban environment. Nevertheless, satisfactory performance is shown for both environments.

• 한국멀티미디어학회논문지
• /
• 제9권11호
• /
• pp.1421-1434
• /
• 2006

공간 데이터 웨어하우스는 SOLAP(Spatial On-Line Analytical Processing)을 이용하여 의사 결정에 필요한 분석 정보를 제공한다. SOLAP은 대용량 데이터를 분석하기 때문에 사전집계를 이용하여 분석비용을 줄이기 위한 많은 연구가 진행되었다. 기존 기법들은 고정크기노드를 갖는 색인을 이용하여 개념계층을 지원하였다. 따라서 산개분포 영역에는 빈 공간이 많이 발생하며, 밀집분포 영역에는 개념계층을 지원할 수 없다. 본 논문은 공간 데이터의 개념계층기반으로 사전집계 색인의 동적 구성 기법을 제안한다. 제안 기법은 트리구조를 이용하여 개념계층의 레벨을 트리의 레벨과 같도록 지원한다. 하나의 노드는 데이터가 적을 경우 엔트리를 분할하여 서로 다른 부모 엔트리를 가질 수 있으며, 데이터가 많을 경우 노드의 연결리스트를 이용하여 같은 레벨에 순차적으로 저장한다. 따라서 데이터가 산개된 분포의 노드에 대해서 저장 공간의 낭비를 최소화하며, 데이터가 밀집한 영역의 노드에 대해서도 노드의 연결리스트로 노드가 분할되지 않으므로 개념계층을 지원할 수 있다. 성능평가를 통하여 색인 구축 시간이 다른 기법과 비슷하고, 색인의 저장 공간이 감소하며, 집계정보의 검색 성능이 다른 기법에 비해 우수한 것을 보인다.