• The KIPS Transactions:PartB
• /
• v.15B no.1
• /
• pp.53-60
• /
• 2008

This paper proposes a new multiobjective optimization method for discriminative training of hidden Markov models (HMMs) used as the recognizer for automatic lipreading. While the conventional Baum-Welch algorithm for training HMMs aims at maximizing the probability of the data of a class from the corresponding HMM, we define a new training criterion composed of two minimization objectives and develop a global optimization method of the criterion based on simulated annealing. The result of a speaker-dependent recognition experiment shows that the proposed method improves performance by the relative error reduction rate of about 8% in comparison to the Baum-Welch algorithm.

• The Journal of Korean Society for Neurotherapy
• /
• v.22 no.3
• /
• pp.25-29
• /
• 2018

Purpose In this study, 45 students were divided into task - oriented group, task - oriented intermittent group, and continuous group, and the effect of each variable on learning was changed to angle of the arms' we checked. Methods This study was conducted for 45 days from March 13 to 18, 2017 for five days in the 20 generals living in Changwon city. Dart and goniometer were used for the measurement. In the exercise program, darts were placed on a smooth, hard surface of 1.5M height and the distance to the subject was set 1.5m away. In the dart throw practice, all three groups were marked with an elbow flexion $30^{\circ}$ on the wall, and the subjects were instructed to throw the dart at that point. Results Comparisons between groups showed significant differences between the continuous task-oriented training group and the intermittent task-oriented training group and the control group at the end of training (*** P <0.001). Conclusion This study showed that the intermittent task-oriented training method showed the highest effect (*** P <0.001), which proved that the intermittent method is the most effective exercise method among the three groups.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.14 no.12
• /
• pp.4664-4681
• /
• 2020

Modulation recognition (MR) plays a key role in cognitive radar, cognitive radio, and some other civilian and military fields. While existing methods can identify the signal modulation type by extracting the signal characteristics, the quality of feature extraction has a serious impact on the recognition results. In this paper, an end-to-end MR method based on long short-term memory (LSTM) and the gated recurrent unit (GRU) is put forward, which can directly predict the modulation type from a sampled signal. Additionally, the sliding window method is applied to fast-changing mixed-modulation signals for which the signal modulation type changes over time. The recognition accuracy on training datasets in different SNR ranges and the proportion of each modulation method in misclassified samples are analyzed, and it is found to be reasonable to select the evenly-distributed and full range of SNR data as the training data. With the improvement of the SNR, the recognition accuracy increases rapidly. When the length of the training dataset increases, the neural network recognition effect is better. The loss function value of the neural network decreases with the increase of the training dataset length, and then tends to be stable. Moreover, when the fast-changing period is less than 20ms, the error rate is as high as 50%. As the fast-changing period is increased to 30ms, the error rates of the GRU and LSTM neural networks are less than 5%.

• International Journal of Contents
• /
• v.17 no.2
• /
• pp.20-31
• /
• 2021

This study aimed to pilot test a newly developed bilateral upper limb rehabilitation training program for improving the upper limb function of individuals with chronic stroke using a visual feedback method. The double-group pretest-posttest design pilot study included 10 individuals with chronic stroke (age >50 years). The intervention (four weekly meetings) consisted of five upper limb training protocols (wrist extension; forearm supination and pronation; elbow extension and shoulder flexion; weight-bearing shift; and shoulder, elbow, and wrist complex movements). Upper limb movement function recovery was assessed with the FuglMeyer Assessment of the Upper Extremity, the Wolf Motor Function Test, the Trunk Control Test, the modified Ashworth Scale, and the visual analog scale at baseline, immediately after, and four weeks after the intervention. The Fatigue Severity Scale was also employed. The Fugl-Meyer Assessment of the Upper Extremity and Wolf Motor Function Test showed significant improvement in upper limb motor function. The Trunk Control Test results increased slightly, and the modified Ashworth Scale decreased slightly, without statistical significance. The visual analog scale scores showed a significant decrease and the Fatigue Severity Scale scores were moderate or low. The bilateral upper limb training program using the visual feedback method could result in slight upper limb function improvements in individuals with chronic stroke.

• The Journal of the Acoustical Society of Korea
• /
• v.41 no.3
• /
• pp.335-341
• /
• 2022

In this paper, we adopt a transformer that shows remarkable performance in natural language processing as an acoustic model of hybrid speech recognition. The transformer acoustic model uses attention structures to process sequential data and shows high performance with low computational cost. This paper proposes a method to improve the performance of transformer AM by applying each of the four algorithms of sequence discriminative training, a weighted finite-state transducer (wFST)-based learning used in the existing DNN-HMM model. In addition, compared to the Cross Entropy (CE) learning method, sequence discriminative method shows 5 % of the relative Word Error Rate (WER).

• The Journal of Korean Academic Society of Nursing Education
• /
• v.14 no.1
• /
• pp.108-116
• /
• 2008

Purpose: The purpose of this study was to identify the experiences of Team-Spirit training in new nurses. Method: Data was collected through open-ended and self-reported questionnaires which were received from 47 new nurses who had finished team-spirit training for 2 days. The content analysis method was used to derive the core-category, categories and concepts of Team-Spirit training for new nurses. Result: The care category identified in new nurses trained in Team-Spirit was upgrade myself and our team. The following 4 categories also emerged; companion, interdependency, importance of community, and future growth of myself and our community. The derived 12 subcategories were intimacy, importance of companions, binding, partaking in difficulties, empowerment, observing rules, cooperation, consensus, self-pledge as a subordinator, motivating vision formation, developing professionalism, contribution and devotion. The 34 concepts were derived from the new nurses' statements. Conclusion: These results imply that Team-Spirit Training for nurses could contribute to companionship, interdependency, importance of community, and future growth of oneself and the hospital team.

• Journal of Korea Multimedia Society
• /
• v.17 no.12
• /
• pp.1446-1452
• /
• 2014

Existing LDA uses the transform matrix that maximizes distance between classes. So we have to convert from an image to one-dimensional vector as training vector. However, in 2D-LDA, we can directly use two-dimensional image itself as training matrix, so that the classification performance can be enhanced about 20% comparing LDA, since the training matrix preserves the spatial information of two-dimensional image. However 2D-LDA uses same calculation schema for transformation matrix and therefore both LDA and 2D-LDA has the heteroscedastic problem which means that the class classification cannot obtain beneficial information of spatial distances of class clusters since LDA uses only data correlation-based covariance matrix of the training data without any reference to distances between classes. In this paper, we propose a new method to apply training matrix of 2D-LDA by using WPS-LDA idea that calculates the reciprocal of distance between classes and apply this weight to between class scatter matrix. The experimental result shows that the discriminating power of proposed 2D-LDA with weighted between class scatter has been improved up to 2% than original 2D-LDA. This method has good performance, especially when the distance between two classes is very close and the dimension of projection axis is low.

• Journal of Korean Academy of Nursing
• /
• v.34 no.2
• /
• pp.213-224
• /
• 2004

Purpose: This study was aimed to evaluate the effect of progressive muscle relaxation training using biofeedback on perceived stress, stress response, immune response and climacteric symptoms, Method: This was a crossover, pre-post test design, The study subjects are 36 middle-aged women who were selected at 2 public health centers, The independent variable was Biofeedback training for 4 weeks, twice a week and home training for 4 weeks, Dependent variables were perceived stress, stress response, immune response, and climacteric symptoms measured with Hildtch's scale (1996), Result: Progressive muscle relaxation training using biofeedback was not effective in reducing perceived stress, but it was shown to be effective in reducing physiological stress responses such as pulse rate and EMG, Though blood pressure and skin conductance were repeatedly down, and skin temperature slowly increased, there were no statistically significant differences. Progressive muscle relaxation training using biofeedback was not effective in reducing serum cortisol, enhancing immune responses, or decreasing climacteric symptoms. Conclusion: The findings point to a pressing need for further, well-controlled and designed research with consideration in selection of subjects and instruments, frequency of measurements, the sampling method, and intervention modalities.

• The Journal of Korean Physical Therapy
• /
• v.23 no.6
• /
• pp.71-76
• /
• 2011

Purpose: We investigated a better method to enhance the vestibular system including balancing by comparing the vestibular stimulation exercise (VSE) and galvanic vestibular stimulation (GVS). Methods: The study was performed with 40 subjects randomized into four groups, including a control group, a VSE group, a GVS group, and a VSE with GVS group. The subjects of VSE performed a forward and backward roll, a right side and left side roll, and an equilibrium board in vestibular stimulation training. GVS was applied for 10 minutes and the cathode and anode side were then changed and GVS was then applied for the remaining 10 minutes. GVS was applied for 20 minutes to the subjects of this group after completion of the VSE program. Results: In the control group, all conditions were significantly decreased (p<0.05) compared to the VSE with GVS group. Also, the center of pressure (CoP) surface was more significantly decreased (p<0.01) and the CoP speed was significantly decreased in the one legged stance (p<0.05) in the control group compared to the GVS group. Conclusion: These findings suggested that GVS training increases balance ability in a narrow width. VSE with GVS training is therefore recommend as the superior method. Using GVS or VSE with GVS training is considered to clinically improve balance ability by stimulating the vestibular system.

• International Journal of Advanced Culture Technology
• /
• v.7 no.4
• /
• pp.23-31
• /
• 2019

This study was a quasi-experimental study with a single group pre-posttest design for evaluating the effects and meanings of a global health training program for nursing students. A mixed data collection method utilizing both qualitative and quantitative methods was applied. The qualitative method consisted of reflection journal entry analysis, while the quantitative study evaluated data collected from a self-reported questionnaire. After attending the program, cultural competencies of participants significantly increased than pre-test while global citizenship was not. Through an analysis of reflection journals written during the overseas training period, the meanings of experiences derived after the program were classified into five themes and 12 sub themes, which were, 'improvement of global health leadership', 'experience of barriers in reality', 'improvement of nursing professional value', 'inspiration of a perspective on health equity', and 'internal growth'. The results indicate that global health training program evaluated in this study was meaningful; it provided students with opportunities to experience the nurse's role in diverse global environments and attempted to measure the effects.