• 한국체육학회지인문사회과학편
• /
• 제53권5호
• /
• pp.521-531
• /
• 2014

본 연구는 한국인 20대의 남성과 여성을 대상으로 보행 분석상의 표준화 방법 중의 하나인 dimensionless number의 효과를 검정하고 이를 통해 성별간 보행형태를 분석하는데 목적이 있다. 피험자는 기술표준원에서 제공하는 한국인 표준체형 및 연령 분류 체계에 맞춰 선정하였으며, 3차원 동작분석 시스템이 사용되었다. 데이터 분석을 위한 소프트웨어로는 Cortex, OrthoTrak, Matlab, Excel이 사용되었으며 통계검정을 위해서 SPSS를 사용하였다. 분석 결과를 살펴보면, Hof(1996)의 dimensionless number 변환을 통한 20대 성별 간 보행 형태는 시·공간 변인인 stride length, step length, stride time, step time, 보행속도(velocity), cadence 모두에서 유의한 차이가 없었으며, 표준화 전·후에 따라 통계분석의 결과가 달라짐을 확인하였다. 따라서 보행 분석에서 데이터의 표준화 방법 중의 하나인 dimensionless number의 적용은 통계학적 검정에 영향을 줄만큼 C.V. 값을 변화시키는 것으로 확인되었다. 본 연구를 통해, 상호 비교를 위한 보행연구에서 dimensionless number를 이용한 표준화 방식은 피험자의 신체적 특성이 분석에 미치는 영향을 제거하고 보다 정확한 통계 검정을 위해서 반드시 요구되는 과정이라는 것을 확인할 수 있었다.

• 한국정보처리학회논문지
• /
• 제3권4호
• /
• pp.883-898
• /
• 1996

본 논문에서는 패턴인식 시스템의 성능 향상을 목적으로 원영상의 데이타량을 압 축하고 난 뒤 형상을 인식하는 개선된 방법을 제안한다. 제안한 방법에서는 수리형태 론적 연산을 사용하여 원영상을 미리 스켈리턴변환하여 데이터 량을 줄이고, 변환된 영상에서 이동 및 크기의 정규화와 회전불변의 과정을 수행하여 패턴을 정합하였다. 크기의 정규화는 형상인식에 필요한 픽셀의 수를 최소로 하여 정합을 하기 위하여 스켈리턴의 픽셀들에 가중치를 부여하고 이를 이용하여 크기를 조정하였다. 따라서 원영상에서 수행하는 이러한 과정들을 스켈리턴 영상에서 수행하게 함으로써 데이터 량이 크게 줄어들게 되어 기억장소의 용량이 최소화되고 연산의 량도 줄어들어 계산의 속도를 고속화 할 수 있게 하였다. 실험을 통하여 인식에 필요한 최적의 크기 인수를 조사하였고, 제안한 방법이 실제의 인식 시스템 구현시 유용하게 사용할 수 있음을 확인할 수 있었다.

• 대한인간공학회:학술대회논문집
• /
• 대한인간공학회 1996년도 추계학술대회논문집
• /
• pp.21-25
• /
• 1996

The Electomyographic (EMG) signals of flexor-extensor muscle pairs were investigated to identify the neural excitation pattern of low-back pain (LBP) patients during a repetitive bending motion. New parameters and EMG normalization technique were developed to quantitatively represent the difference of temporal EMG patterns between ten healthy subjects and ten LBP patients. Flexor-extensor muscle pairs such as rectus abdominis(RA)-erector spinae (ES at LS), external oblique(EO)-internal oblique(IO), rectus femois (quadriceps: QUD)-biceps femoris(hamstrings:HAM), and tibialis anterior(TA)-gastrocnemius(GAS) pairs of muscles were selected in this study. Results indicated that the temporal EMG pattern such as the peak timing difference of QUD-HAM muscle pair and the duration of coexcitation of ES-RA muscle pair showed a statistically isgnificant difference between healthy subjects and LBP patients. These results indicated that the new technique and parameters could be used as a diagnostic tool especially for LBP patients with soft tissue injuries that are rarely dentified by traditional imaging techniques such as X-ray, CT scan or MRI. Improtantly, the new EMG technique did not require the maximal volutary contraction(MVC) measure for normalization that helped patients minimize the pain experience during and after the session. Further study needs to be made to validate and refine this method for clinical application.

• Structural Engineering and Mechanics
• /
• 제84권2호
• /
• pp.181-206
• /
• 2022

This paper presents experimental modal analysis and static load testing results to validate the accuracy of dynamic parameters-based damage locating indices in RC structures. The study investigates the accuracy of different dynamic-based damage locating indices compared to observed crack patterns from static load tests and how different damage levels and scenarios impact them. The damage locating indices based on mode shape curvature and mode shape fourth derivate in their original forms were found to show anomalies along the beam length and at the supports. The modified forms of these indices show higher sensitivity in locating single and multi-cracks at different damage scenarios. The proposed stiffness reduction index shows good sensitivity in detecting single and multi-cracks. The proposed anomalies elimination procedure helps to remove the anomalies along the beam length. Also, the adoption of the proposed weighting method averaging procedure and normalization procedure help to draw the overall crack pattern based on the adopted set of modes.

• 대한전기학회:학술대회논문집
• /
• 대한전기학회 1992년도 하계학술대회 논문집 A
• /
• pp.462-464
• /
• 1992

In this paper, an extraction of depth intonation in pattern using neural network is presented. All the 3D images represent the depth information in grey pixels. This pixels which have analog values translated digital values. Because of the noise and distortion in pattern, we use the normalization in learning and recalling the patterns. Our method has eight direction vectors and slopes for pattern. Also, we use potential to obtain the mean slope and direction vectors of given 3D patches. The higher level of deduction finding the global depth information is also carried out by using neural network.

• E2M - 전기 전자와 첨단 소재
• /
• 제11권11호
• /
• pp.1-8
• /
• 1998

This paper presents an arificial neuro-fuzzy technique based prtial discharge (PD) pattern classifier to power system application. This may require a complicated analysis method employ -ing an experts system due to very complex progressing discharge form under exter-nal stress. After referring briefly to the developments of artificical neural network based PD measurements, the paper outlines how the introduction of new emerging technology has resulted in the design of a number of PD diagnostic systems for practical applicaton of residual lifetime prediction. The appropriate PD data base structure and selection of learning data size of PD pattern based on fractal dimentsional and 3-D PD-normalization, extraction of relevant characteristic fea-ture of PD recognition are discussed. Some practical aspects encountered with unknown stress in the neuro-fuzzy techniques based real time PD recognition are also addressed.

• 한국정보통신학회논문지
• /
• 제25권11호
• /
• pp.1649-1654
• /
• 2021

본 논문에서는 고령자의 낙상상황을 감지할 수 있는 텐서플로우 장단기 메모리 기반 낙상감지 시스템의 정규화에 대하여 소개한다. 낙상감지는 고령자의 몸에 부착한 3축 가속도 센서 데이터를 사용하며, 총 7가지의 행동 패턴들에 대하여 학습하며, 각각 4가지는 일상생활에서 일어나는 패턴이고, 나머지 3가지는 낙상에 대한 패턴이다. 학습시에는 손실함수(loss function)를 효과적으로 줄이기 위하여 정규화 과정을 진행하며, 정규화 과정은 데이터에 대하여 최대최소 정규화, 손실함수에 대하여 L2 정규화 과정을 진행한다. 3축 가속도 센서를 이용하여 구한 다양한 파라미터에 대하여 정규화 과정의 최적의 조건을 제시한다. 낙상 검출율면에서 SVM을 이용하고 정규화 127과 정규화율 λ 0.00015일 때 Sensitivity 98.4%, Specificity 94.8%, Accuracy 96.9%로 가장 좋은 모습을 보였다.

• 대한인간공학회지
• /
• 제14권2호
• /
• pp.1-14
• /
• 1995

Neuromuscular difference between normal subjects and low-back pain patients has been identified in terms of neural excitation signal measured by Electromyography (EMG) under the dynamic flexion/extension trunk motion. Ten healthy subjects and ten low-back pain patients were recruited for this study. New parameters and normalization technique were introduced to quantify the muscle excitation pattern among the flexor-extensor pairs of muscles : rectus abdominis (RA)-erector spinae (ES at L1 and L5 level), external oblique (EO)-internal oblique (IO), rectus femoris (quadricep : QUD)-biceps femoris( hamstring : HAM), and tibialis anterior (TA)-gastrocnemius (GAS). Results indicated that the temporal EMG pattern such as peak timing difference between the hip flexor (QUD) and extensor (HAM) and the duration of coexcitation between ES at L5 and RA muscle pairs showed a statistically significant difference between normal subjects and low-back pain patients. Improtantly, this study presented a new technique to identify the dynamic muscle excitation pattern that canb be least affected by EMG-length-velocity relationship. Further study can performed to validate this method for clinical application to quantitatively identify the low-back pain patients in the future.

• 전기학회논문지
• /
• 제67권1호
• /
• pp.114-123
• /
• 2018

This paper concerns a design and learning method of softmax function neural networks based on K-means clustering. The partial discharge data Information is preliminarily processed through simulation using an Epoxy Mica Coupling sensor and an internal Phase Resolved Partial Discharge Analysis algorithm. The obtained information is processed according to the characteristics of the pattern using a Motor Insulation Monitoring System program. At this time, the processed data are total 4 types that void discharge, corona discharge, surface discharge and slot discharge. The partial discharge data with high dimensional input variables are secondarily processed by principal component analysis method and reduced with keeping the characteristics of pattern as low dimensional input variables. And therefore, the pattern classifier processing speed exhibits improved effects. In addition, in the process of extracting the partial discharge data through the MIMS program, the magnitude of amplitude is divided into the maximum value and the average value, and two pattern characteristics are set and compared and analyzed. In the first half of the proposed partial discharge pattern classifier, the input and hidden layers are classified by using the K-means clustering method and the output of the hidden layer is obtained. In the latter part, the cross entropy error function is used for parameter learning between the hidden layer and the output layer. The final output layer is output as a normalized probability value between 0 and 1 using the softmax function. The advantage of using the softmax function is that it allows access and application of multiple class problems and stochastic interpretation. First of all, there is an advantage that one output value affects the remaining output value and its accompanying learning is accelerated. Also, to solve the overfitting problem, L2-normalization is applied. To prove the superiority of the proposed pattern classifier, we compare and analyze the classification rate with conventional radial basis function neural networks.

• 제어로봇시스템학회:학술대회논문집
• /
• 제어로봇시스템학회 1990년도 한국자동제어학술회의논문집(국제학술편); KOEX, Seoul; 26-27 Oct. 1990
• /
• pp.939-944
• /
• 1990

To recognize isomorphic transformation patterns, such as scale-change, translation and rotation transformed patterns, is an old difficult but interesting problem. Many researches have been done with a dominant approach of normalization by many eminent pioneers. However, there seems no a perfect system which can even recognize 90 .deg.-multiple rotation isomorphic transformation patterns for real needs. Here, as a new challenge, we propose a method of how to recognize 90 .deg.-multiple rotation isomorphic and symmetry isomorphic transformation patterns.