• 전자공학회논문지
• /
• 제51권5호
• /
• pp.197-206
• /
• 2014

운동심상(Motor imagery) 기반의 뇌-컴퓨터 인터페이스(Brain-computer Interface)는 주로 뇌전도(Electroencephalography, EEG)를 이용하여 사용자의 자발적인 운동 의지를 읽는 기술로 최근 주목받고 있다. 이 중에서도 피실험자의 운동 의지를 정확히 해석하기 위해 감각운동 영역(sensorimotor area)의 일부분에서 나타나는 ${\mu}$-대역(8-13Hz)의 전위 감소 현상인 event related desynchronization(ERD)을 분석하는 연구가 많이 진행되고 있다. 하지만 EEG는 공간 해상도가 낮고 사용자에 따라 ERD가 발생하는 주파수 대역이 다소 차이가 있어 추정에 어려움이 있다. 이에 대한 개선 방법의 하나로서 공간 필터를 구현하는 common spatial pattern (CSP)과 필터 뱅크(filter bank)를 결합한 형태인 discriminative filter bank common spatial pattern(DFBCSP)이 제안되었다. 그러나 DFBCSP는 EEG 신호의 평균 전력(power)의 Fisher ratio를 이용하여 사용자에 따른 효과적인 주파수 대역을 포함하는 discriminative filter bank(DFB)를 구성하여 분류 정확도를 향상시켰지만 ERD의 공간 패턴이 나타나는 적절한 필터를 선택하지 않는 경우가 발생한다. 이러한 문제를 해결하기 위해 본 논문에서는 EEG 신호의 평균전력 대신 CSP의 특성 벡터를 이용하여 DFB를 구성하는 방법을 제안한다. 기존의 방법과 제안한 방법의 필터 선택 결과와 분류 정확도 분석을 통해 CSP 특성 벡터가 DFB 구성에 더욱 효과적임을 보인다.

• 한국HCI학회:학술대회논문집
• /
• 한국HCI학회 2009년도 학술대회
• /
• pp.642-645
• /
• 2009

Brain Computer Interface (BCI) is a communication pathway between devices (computers) and human brain. It treats brain signals in real-time basis and discriminates some information of what human brain is doing. In this work, we develop a EEG BCI system using a feature extraction such as common spatial pattern (CSP) and a classifier using Fisher linear discriminant analysis (FLDA). Two-class EEG motor imagery movement datasets with both cued and uncued are tested to verify its feasibility.

• 한국정보과학회:학술대회논문집
• /
• 한국정보과학회 2011년도 한국컴퓨터종합학술대회논문집 Vol.38 No.1(A)
• /
• pp.142-145
• /
• 2011

뇌전도 기반 뇌-컴퓨터 인터페이스 기술은 신체 움직임이 불가능하거나 불편한 사람에게 새로운 의사전달 수단이 될 수 있으며 일반인에게도 상상만으로 컴퓨터 혹은 기계에 명령을 내릴 수 있게 하는 기술이다. 본 논문에서는 뇌-컴퓨터 인터페이스 연구 분야에 잘 알려진 Common Spatial Pattern (CSP), Invariant Common Spatial Pattern (iCSP) 그리고 Common Spatio-Spectral Pattern (CSSP) 알고리즘들의 성능을 비교 분석하였고, CSSP에 불변성(invariant)을 고려한 iCSSP를 제안하였다. 9명의 피험자로부터 상상움직임 실험을 통해 18셋의 뇌전도 데이터를 측정하였고, 4가지 알고리즘들을 성능 면에서 비교하였다. 그 결과 CSSP의 성능과 차이가 크지는 않지만, 본 연구에서 제안한 노이즈를 고려하여 최적의 필터를 구성하는 iCSSP에 대하여 더 나은 성능을 보여주는 결과들을 확인할 수 있었다.

• Ocean and Polar Research
• /
• 제29권4호
• /
• pp.391-400
• /
• 2007

To identify the spatial distribution pattern of water quality in Masan Bay, Pearson's correlation as a common statistic method and Moran's I as a spatial autocorrelation statistics were applied to the hydrological data seasonally collected from Masan Bay for two years ($2004{\sim}2005$). Spatial distribution of salinity, DO and silicate among the hydrological parameters clustered strongly while chlorophyll a distribution displayed a weak clustering. When the similarity matrix of Moran's I was compared with correlation matrix of Pearson's r, only the relationships of temperature vs. salinity, temperature vs. silicate and silicate vs. total inorganic nitrogen showed significant correlation and similarity of spatial clustered pattern. Considering Pearson's correlation and the spatial autocorrelation results, water quality distribution patterns of Masan Bay were conceptually simplified into four types. Based on the simplified types, Moran's I and Pearson's r were compared respectively with spatial distribution maps on salinity and silicate with a strong clustered pattern, and with chlorophyll a having no clustered pattern. According to these test results, spatial distribution of the water quality in Masan Bay could be summed up in four patterns. This summation should be developed as spatial index to be linked with pollutant and ecological indicators for coastal health assessment.

• 산업공학
• /
• 제24권4호
• /
• pp.447-456
• /
• 2011

Identifying the major moving patterns of shoppers' movements in the selling floor has been a longstanding issue in the retailing industry. With the advent of RFID technology, it has been easier to collect the moving data for a individual shopper's movement. Most of the previous studies used the traditional clustering technique to identify the major moving pattern of customers. However, in using clustering technique, due to the spatial constraint (aisle layout or other physical obstructions in the store), standard clustering methods are not feasible for moving data like shopping path should be adjusted for the analysis in advance, which is time-consuming and causes data distortion. To alleviate this problems, we propose a new approach to spatial pattern clustering based on longest common subsequence (LCSS). Experimental results using the real data obtained from a grocery in Seoul show that the proposed method performs well in finding the hot spot and dead spot as well as in finding the major path patterns of customer movements.

• 수산해양교육연구
• /
• 제23권3호
• /
• pp.477-485
• /
• 2011

For the identifying of spatial distribution pattern, Moran's Index(I) which has the range of values from -1 to +1 is common method for the spatial autocorrelation measurement. When I is close to 1, all neighboring features have close to the same value, indicating clustered pattern. Conversely, if the spatial pattern is dispersed, I is close to -1. And I closing to 0 means spatially random pattern. However, this index equation is influenced by how defining the neighboring features for target feature. To compare and understand the difference of neighborhood definition methods, fixed distance neighboring method and Gabriel Network method were used for I. In this study, these two methods were applied to two marine environments with water quality data. One is Gwangyang Bay which has complex geometric coastal structure located in South Sea of Korea. Another is Uljin area adjacent to open sea located in east coast of Korea. The distances between water quality observed locations were relatively regular in Gwangyang Bay, however, irregular in Uljin area. And for the fixed distance method popular Arc GIS tool was used, but, for the Gabriel Network, Visual Basic program was developed to produce Gabriel Network and calculate Moran's I and its Z-score automatically. According to this experimental results, different spatial pattern was showed differently for some data with using of neighboring definition methods. Therefore there is need to choose neighboring definition method carefully for spatial pattern analysis.

• International Journal of Fuzzy Logic and Intelligent Systems
• /
• 제15권4호
• /
• pp.277-282
• /
• 2015

Braincomputer interface (BCI) technology is making advances in the field of humancomputer interaction (HCI). To improve the BCI technology, we study the changes in the electroencephalogram (EEG) signals for six levels of grip strength: 10%, 20%, 40%, 50%, 70%, and 80% of the maximum voluntary contraction (MVC). The measured EEG data are categorized into three classes: Weak, Medium, and Strong. Features are then extracted using power spectrum analysis and multiclass-common spatial pattern (multiclass-CSP). Feature datasets are classified using a support vector machine (SVM). The accuracy rate is higher for the Strong class than the other classes.

• 한국정보처리학회:학술대회논문집
• /
• 한국정보처리학회 2016년도 추계학술발표대회
• /
• pp.569-570
• /
• 2016

본 논문은 뇌파 기반 감정 분류 기술의 상용화를 위한 낮은 성능을 보완하기 위하여 Regularized Common Spatial Pattern 필터링을 통한 전처리 방식을 제안하고 있다. RCSP 필터는 뇌파 기반 행동 인식 시스템에서 높은 성능 향상을 보이는 것으로 알려져 있다. 본 연구에서는 장기적이고 복합적인 뇌파의 감성 인지 연구에도 RCSP 필터의 적용 방법을 설명하고, 제안하는 알고리즘이 뇌파를 통한 감정 인식에 성능 향상을 보여준다는 것을 설명하고 있다.

• IEIE Transactions on Smart Processing and Computing
• /
• 제4권2호
• /
• pp.97-102
• /
• 2015

In this paper, an extension of the standard common spatial pattern (CSP) algorithm using the strong uncorrelated transform (SUT) is used in order to extract the features for an accurate classification of the left- and right-hand motor imagery tasks. The algorithm is designed to analyze the complex data, which can preserve the additional information of the relationship between the two electroencephalogram (EEG) data from distant channels. This is based on the fact that distant regions of the brain are spatially distributed spatially and related, as in a network. The real-world left- and right-hand motor imagery EEG data was acquired through the Physionet database and the support vector machine (SVM) was used as a classifier to test the proposed method. The results showed that extracting the features of the pair-wise channel data using the strong uncorrelated transform complex common spatial pattern (SUTCCSP) provides a higher classification rate compared to the standard CSP algorithm.

• 한국지능시스템학회논문지
• /
• 제25권1호
• /
• pp.29-34
• /
• 2015

Brain-computer interface 기술은 일상에서 편안한 생활을 위해 다방면으로 연구가 진행 중이다. 본 연구는 3가지 동작의 뇌파특성을 분석하여 의수와 같은 외부기기의 세밀한 동작 제어를 목적으로 한다. 피험자들은 악력기를 쥘 때 (Grip), 손가락만을 움직일 때 (Move), 아무런 동작을 취하지 않을 때 (Relax)의 3가지 동작을 수행하였고, 뇌파를 측정하여 power spectrum analysis와 multi-common spatial pattern 알고리즘으로 특징추출을 수행하였으며, 분류알고리즘인 SVM(support vector machine)으로 뇌파의 특징데이터들을 분류하였다. 실험결과 3개의 다른 동작을 분류한 결과, 실험에 참여한 3명의 피험자 중 2명에게서 Grip 클래스의 분류율이 가장 높은 분류율을 보였다. 본 연구의 결과는 뇌파를 이용하여 의수가 필요한 환자들에게 유용할 것으로 기대한다.