• International Journal of Fuzzy Logic and Intelligent Systems
• /
• 제4권2호
• /
• pp.236-240
• /
• 2004

Stroke-based composite HMMs with articulation states are proposed to deal with 3D spatio-temporal trajectory gestures. The direct use of 3D data provides more naturalness in generating gestures, thereby avoiding some of the constraints usually imposed to prevent performance degradation when trajectory data are projected into a specific 2D plane. Also, the decomposition of gestures into more primitive strokes is quite attractive, since reversely concatenating stroke-based HMMs makes it possible to construct a new set of gesture HMMs without retraining their parameters. Any deterioration in performance arising from decomposition can be remedied by a partial tuning process for such composite HMMs.

• Journal of Platform Technology
• /
• 제6권1호
• /
• pp.3-8
• /
• 2018

This paper proposes a technique for recognizing online handwritten cursive data obtained by tracing a motion trajectory while a user is in the 3D space based on a convolution neural network (CNN) algorithm. There is a difficulty in recognizing the virtual character input by the user in the 3D space because it includes both the character stroke and the movement stroke. In this paper, we divide syllable into consonant and vowel units by using labeling technique in addition to the result of localizing letter stroke and movement stroke in the previous study. The coordinate information of the separated consonants and vowels are converted into image data, and Korean handwriting recognition was performed using a convolutional neural network. After learning the neural network using 1,680 syllables written by five hand writers, the accuracy is calculated by using the new hand writers who did not participate in the writing of training data. The accuracy of phoneme-based recognition is 98.9% based on convolutional neural network. The proposed method has the advantage of drastically reducing learning data compared to syllable-based learning.

• 한국운동역학회지
• /
• 제26권4호
• /
• pp.333-341
• /
• 2016

Objective: The purpose of this study was to assess the consistency of the gliding and push-off motion for single leg skating from the first to fourteenth steps. We hypothesized that: 1) there would be no difference in stroke trajectory, step rate, and cycle rate between the left and right steps of gliding; and 2) there would be a difference in the resultant velocity of toe push-off and the horizontal velocity of the center of mass after six step push-offs. Method: The study included five male 500-m speed skaters (mean height, $1.80{\pm}0.02m$; mean weight, $76.8{\pm}3.96kg$; record, $35.83{\pm}0.30sec$; 100-m record, <9.97 sec). Data were collected from the first to fourteenth steps (40 m) and recorded using five digital JVC GR-HD1KR video cameras (Victor Co., Japan) operating at a sampling frequency of 60 fields/sec and shutter speed of 1/500 sec. For each film frame, the joint positions were digitized using the KWON3D motion analyzer. Position data were filtered with low-pass Butterworth $4^{th}$ order at the cut-off frequency of 7.4 Hz. Results: The right toe of the skating trajectories at $2^{nd}$, $5^{th}$, and $7^{th}$ strokes differed from those of the left toe. The angles of the right and left knee demonstrated unbalanced patterns from the flexion and extension legs. The step and cycle rates of the right and left leg differed from the start until 20 m. The resultant velocities of the toe at the push-off phase and of the body mass center diverged before the six push-offs. Conclusion: This study's findings indicate that the toe of skating trajectory on left and right sliding after push-off should maintain a symmetrical trajectory. The resultant velocity of toe push-off and horizontal velocity from the center of body need to be separated after about six step push-offs.

• 전자공학회논문지C
• /
• 제36C권8호
• /
• pp.46-53
• /
• 1999

본 논문에서는 원격 작업 환경에서 명령자의 제스처를 자동으로 인식하기 위하여 글로브 센서 기반의 인식 기법을 이용한 3차원 제스처 인식 시스템을 구현하였다. 인식 시스템은 PinchGlove에 부착하여 사용되는 Polhemus 센서로부터 획득한 손 궤적의 3차원 위치 좌표열을 입력으로 사용한다. 또한 본 논문에서는 제스처 인식을 위해 제스처를 구성하는 스트로크를 기본 인식 단위로 사용하는 방법을 제안한다. 각 스트로크는 이산 HMM으로 모델링 되며 이들 HMM을 연결하여 생성된 결합 HMM으로 원격 작업에 사용될 각 제스처들을 모델링 한다. 이 방법은 새로이 정의되는 제스처에 대해 추가의 학습 과정을 필요로 하지 않아 인식 시스템의 확장성을 높일 수 있다. 16개의 제스처를 사용한 인식 실험에서 스트로크 기반의 결합 HMM은 제스처를 기본 단위로 사용한 HMM에 비해 더 좋은 인식 성능을 나타냄을 확인할 수 있었다.

• 문화기술의 융합
• /
• 제5권1호
• /
• pp.417-428
• /
• 2019

본 연구는 급성 뇌졸중 입원 환자의 연하장애 유병률과 그 관련요인을 탐색하기 위한 서술적 조사연구로, 일상급종합병원에 급성 뇌졸중으로 진단받은 입원 환자 131명(연령 범위 34-92세, 58.0% 남성)을 대상으로 수행되었다. 본 연구의 결과 급성 뇌졸중 입원 환자의 유병률은 24.4%였으며, 대상자의 나이, 뇌졸중 중증도, 구강상태, 일상활동 의존도가 연하장애에 영향을 미치는 것으로 나타났다. 이러한 결과를 기반으로 뇌졸중 발병 직후부터 연하장애 조기 발견을 위한 적극적인 간호사정과 예방적 중재 프로그램 개발과 적용이 필요하다.

• 전기전자학회논문지
• /
• 제11권1호통권20호
• /
• pp.46-53
• /
• 2007

Design characteristics of PMLSM(Permanent Magnet Linear Synchronous Motor) considering the dynamic running condition under the limited input voltage and current for short reciprocating trajectory are presented. Particularly, the dynamic constraints resulted from the dynamic capability of PMLSM and the required motional performance of the repeated short stroke are applied to determine the design specification of PMLSM. In addition, optimal design flow based on the dynamic constraints is specified with the design parameters, such as coil resistances, the EMF constants, inductances, pole-pitch. Furthermore, proposed methods and results are validated by the experimental ones measured with the purpose-built prototype.

• 제어로봇시스템학회논문지
• /
• 제16권10호
• /
• pp.919-926
• /
• 2010

This paper proposes a wearable hand rehabilitation device, DULEX, for persons with functional paralysis of upper-limbs after stoke. DULEX has three degrees of freedom for rehabilitation exercises for wrist and fingers except the thumb. The main function of DULEX is to extend the range of motions of finger and wrist being contracture. DULEX is designed by using a parallel mechanism, and its parameters such as length and location of links are determined by kinematic analysis. The motion trajectory of the designed DULEX is aligned to human hand to prevent a slip. To reduce total weight of DULEX, artificial air muscles are used for actuating each joint motion. In feedback control, each joint angle is indirectly estimated from the relations of the input air pressure and the output muscle length. Experimental results show that DULEX is feasible in hand rehabilitation for stroke survivals.

• Advances in robotics research
• /
• 제2권3호
• /
• pp.237-245
• /
• 2018

This study evaluates the efficacy of a class robust control scheme namely active force control in performing a joint based trajectory tracking of an upper limb exoskeleton in rehabilitating the elbow joint. The plant of the exoskeleton system is obtained via system identification method whilst the PD gains were tuned heuristically. The estimated inertial parameter that enables the AFC disturbance rejection effect is attained by means of a non-nature based metaheuristic optimisation technique known as simulated Kalman filter (SKF). It was demonstrated from the present investigation that the proposed PDAFC scheme outperformed the classical PD algorithm in tracking the prescribed trajectory both in the presence and without the presence of disturbance attributed by the mannequin limb weights (1 kg and 1.5 kg) that mimics the weight of actual human limb weight. Therefore, it is apparent from the results obtained from the present study that the proposed control scheme, i.e., PDAFC is suitable for the application of exoskeleton for stroke rehabilitation.

• 한국운동역학회지
• /
• 제23권4호
• /
• pp.319-326
• /
• 2013

The objective of this study is to help golfers to select adequate putters and to provide golfers basic scientific data for improving athletic performance by showing differences of kinematic variables according to the shape of putter head. In this research three right-handed male pro-golfer who are listed at KPGA were studied and three video camera (GR-HD1KR, JVC, Japan) were used and recording speed was 60 frame/sec during the research. In this study kinematic variables were calculated using Kwon3D XP program and analysed on the 4 events and 3 phases. This study showed the following results : (1) The swing of heel-toe putter showed longer time than that of face balanced putter, and there are statistical significant difference of kinematic variables of each objective (2) As of the trajectory of putter head, heel-toe putter showed more approximating curve trajectory than that of in-between, face balanced putter on the X-axis (3) heel-toe putter showed longer distance follow-through than that of other putters by statistically significant difference on the Y-axis (4) Heel-toe putter showed longer distance swing trajectory over the ground than that of other putters by statistically significant difference on the Z-axis.

• International Journal of Contents
• /
• 제11권3호
• /
• pp.7-13
• /
• 2015

One of the biggest problems of skeletonization is the occurrence of distortions at the junction point of the final binary image. At the junction area, a single point usually becomes a small stroke, and the corresponding trajectory task, as well as the OCR, consequently becomes more complicated. We therefore propose an adaptive post-processing method that uses an adaptive threshold technique to correct the distortions. Our proposed method transforms the distorted segments into a single point so that they are as similar to the original image as possible, and this improves the static handwriting images after the skeletonization process. Further, we attained promising results regarding the usage of the enhanced skeletonized images in other applications, thereby proving the expediency and efficiency of the proposed method.