• Proceedings of the Korean Information Science Society Conference
• /
• 한국정보과학회 2004년도 가을 학술발표논문집 Vol.31 No.2 (1)
• /
• pp.802-804
• /
• 2004

지적재산권이 산업의 핵심으로 자리잡음으로써 특허의 중요성이 날로 증가하고 있다. 현재 특허문서 검색을 서비스하고 있는 상용시스템의 경우 문서간의 유사도나, 질의어에 따른 순위(Ranking)가 매겨지지 않는 불리언 모델이 검색에 사용되고 있다. 본 논문에서는 유사도에 기반 한 순위화가 가능한 벡터모델기반의 특허검색시스템을 개발하고 시계분야의 시소러스를 구축하여 시계분야의 특허검색 시스템에 적용하였다. 쿼리확장의 성능을 평가하기 위해 10개의 쿼리로 실험하였고 평균 36.2%의 정확도가 향상되었다. 그리고 검색결과의 오른쪽에 시소러스를 제시함으로써 특허검색시스템을 이용하는 사용자에게 추가 질의어를 쉴게 선택할 수 있도록 하여 인터페이스 부분의 향상을 추구하였다.

• Annual Conference on Human and Language Technology
• /
• 한국정보과학회언어공학연구회 2019년도 제31회 한글 및 한국어 정보처리 학술대회
• /
• pp.326-329
• /
• 2019

기계독해를 실용화하기 위해 단락을 검색하는 검색 모델은 최근 기계독해 모델이 우수한 성능을 보임에 따라 그 필요성이 더 부각되고 있다. 그러나 기존 검색 모델은 질의와 단락의 어휘 일치도나 유사도만을 계산하므로, 기계독해에 필요한 질의 어휘의 문맥에 해당하는 단락 검색을 하지 못하는 문제가 있다. 본 논문에서는 이러한 문제를 해결하기 위해 Word2vec의 입력 단어열의 벡터에 해당하는 IN Weight Matrix와 출력 단어열의 벡터에 해당하는 OUT Weight Matrix를 사용한 단락 검색 모델을 제안한다. 제안 방법은 기존 검색 모델에 비해 정확도를 측정하는 Precision@k에서 좋은 성능을 보였다.

• Journal of the Korea Society of Computer and Information
• /
• 제19권9호
• /
• pp.125-139
• /
• 2014

In this paper, I propose a response modeling with a Semi-Supervised Support Vector Regression (SS-SVR) algorithm. In order to increase the accuracy and profit of response modeling, unlabeled data in the customer dataset are used with the labeled data during training. The proposed SS-SVR algorithm is designed to be a batch learning to reduce the training complexity. The label distributions of unlabeled data are estimated in order to consider the uncertainty of labeling. Then, multiple training data are generated from the unlabeled data and their estimated label distributions with oversampling to construct the training dataset with the labeled data. Finally, a data selection algorithm, Expected Margin based Pattern Selection (EMPS), is employed to reduce the training complexity. The experimental results conducted on a real-world marketing dataset showed that the proposed response modeling method trained efficiently, and improved the accuracy and the expected profit.

• Journal of Broadcast Engineering
• /
• 제16권6호
• /
• pp.1069-1072
• /
• 2011

This paper proposes the method which reduces complexity and improves coding efficiency by solving a problem of HEVC bi-prediction. In current HM 3.0, it is frequently occurred that L0 motion information and L1 motion information are identical in blocks which are bi-predicted. In this case, L1 motion vector is replaced by non-zero motion vector which belongs to first available neighbor block of current block. If they are still identical, prediction mode is replaced by uni-prediction. As an experimental result, in LD(Low-Delay) case, decoding time is reduced roughly 2%~5% and coding gain is roughly 0.3%~0.5% compared with the HM 3.0 anchor.

• The Journal of the Acoustical Society of Korea
• /
• 제30권8호
• /
• pp.461-467
• /
• 2011

Accurate classification of input signals is the key prerequisite for variable bit-rate coding, which has been introduced in order to effectively utilize limited communication bandwidth. Especially, recent surge of multimedia services elevate the importance of speech/music classification. Among many speech/music classifier, the ones based on support vector machine (SVM) have a strong selling point, high classification accuracy, but their computational complexity and memory requirement hinder their way into actual implementations. Therefore, techniques that reduce the computational complexity and the memory requirement is inevitable, particularly for embedded systems. We first analyze implementation of an SVM-based classifier on embedded systems in terms of execution time and energy consumption, and then propose two techniques that alleviate the implementation requirements: One is a technique that removes support vectors that have insignificant contribution to the final classification, and the other is to skip processing some of input signals by virtue of strong correlations in speech/music frames. These are post-processing techniques that can work with any other optimization techniques applied during the training phase of SVM. With experiments, we validate the proposed algorithms from the perspectives of classification accuracy, execution time, and energy consumption.

• Journal of Broadcast Engineering
• /
• 제21권4호
• /
• pp.578-591
• /
• 2016

In this paper, the fast encoding algorithm in High Efficiency Video Coding (HEVC) encoder was studied. For the encoding efficiency, the current HEVC reference software is divided the input image into Coding Tree Unit (CTU). then, it should be re-divided into CU up to maximum depth in form of quad-tree for RDO (Rate-Distortion Optimization) in encoding precess. But, it is one of the reason why complexity is high in the encoding precess. In this paper, to reduce the high complexity in the encoding process, it proposed the method by determining the maximum depth of the CU using a hierarchical clustering at the pre-processing. The hierarchical clustering results represented an average combination of motion vectors (MV) on neighboring blocks. Experimental results showed that the proposed method could achieve an average of 16% time saving with minimal BD-rate loss at 1080p video resolution. When combined the previous fast algorithm, the proposed method could achieve an average 45.13% time saving with 1.84% BD-rate loss.

• The Journal of the Acoustical Society of Korea
• /
• 제22권4호
• /
• pp.308-314
• /
• 2003

Wideband speech communications with 7㎑ bandwidth can provide high-quality speech services that are almost impossible with current narrow-band speech communications with 3.4 ㎑ bandwidth, and AMR wideband codec was recently developed for these services. The performance of AMR wideband codec is excellent due to its wideband information and partially to ACELP structure, but it requires high computational complexity especially in codebook search. In this paper, to solve this problem, an efficient codebook search method for AMR wideband codec is proposed. The proposed method first determines the coarse initial codevector, then improves the performance of codevector by replacing a poor pulse in codevector with better one iteratively. Simulations show that AMR wideband codec with proposed codebook search method has higher performance with much less computational cost than conventional AMR wideband codec.

• Journal of the Korea Institute of Information and Communication Engineering
• /
• 제21권5호
• /
• pp.960-966
• /
• 2017

In this paper, we proposed an algorithm that can extract lesion by inputting a medical image. Feature points are extracted using SIFT algorithm to extract genetic training of medical image. To increase the intensity of the feature points, the input image and that raining image are matched using vector similarity and the lesion is extracted. The vector similarity match can quickly lead to lesions. Since the direction vector is generated from the local feature point pair, the direction itself only shows the local feature, but it has the advantage of comparing the similarity between the other vectors existing between the two images and expanding to the global feature. The experimental results show that the lesion matching error rate is 1.02% and the processing speed is improved by about 40% compared to the case of not using the feature point intensity information.

• The Journal of Korean Institute of Communications and Information Sciences
• /
• 제33권6A호
• /
• pp.577-582
• /
• 2008

In this paper, we present a scheme to overcome the worst case complexity of the sphere decoder. If the number of visited nodes reaches the threshold, the detected symbol vector is determined between two candidate symbol vectors. One candidate symbol vector is obtained from the demodulated output of ZF receiver which is initial stage of the sphere decoder. The other candidate symbol vector consists of two sub-symbol vectors. The first sub-symbol vector consists of lately visited nodes running from the most upper layer. The second one contains corresponding demodulated outputs of ZF receiver. Between these two candidate symbol vectors, the one with smaller euclidean distance to the received symbol vector is chosen as detected symbol vector. In addition, we show the upper bound of symbol error rate performance for the sphere decoder using the proposed scheme. In the simulation, the proposed scheme shows the significant reduction of the worst case complexity while having negligible SER performance degradation.

• Journal of KIISE:Software and Applications
• /
• 제30권11호
• /
• pp.1072-1079
• /
• 2003

This study is purported for suggesting a new clustering algorithm that enables incremental categorization of numerous documents. The suggested algorithm adopts the natures of the spherical k-means algorithm, which clusters a mass amount of high-dimensional documents, and the fuzzy ART(adaptive resonance theory) neural network, which performs clustering incrementally. In short, the suggested algorithm is a combination of the spherical k-means vector space model and concept vector and fuzzy ART vigilance parameter. The new algorithm not only supports incremental clustering and automatically sets the appropriate number of clusters, but also solves the current problems of overfitting caused by outlier and noise. Additionally, concerning the objective function value, which measures the cluster's coherence that is used to evaluate the quality of produced clusters, tests on the CLASSIC3 data set showed that the newly suggested algorithm works better than the spherical k-means by 8.04% in average.