• The Journal of the Acoustical Society of Korea
• /
• v.32 no.5
• /
• pp.438-445
• /
• 2013

In this paper, we propose an enhancement of Japanese acoustic model which is trained with Korean speech database by using several combination strategies. We describe the strategies for training more than two language combination, which are Cross-Language Transfer, Cross-Language Adaptation, and Data Pooling Approach. We simulated those strategies and found a proper method for our current Japanese database. Existing combination strategies are generally verified for under-resourced Language environments, but when the speech database is not fully under-resourced, those strategies have been confirmed inappropriate. We made tyied-list with only object-language on Data Pooling Approach training process. As the result, we found the ERR of the acoustic model to be 12.8 %.

• Journal of International Academy of Physical Therapy Research
• /
• v.11 no.2
• /
• pp.2021-2027
• /
• 2020

Background: Previous robot-mediated gait training has been proven several limitations such as pointless repeated motion training, decreased presence, etc. In this research, adult stroke patients were participated in robot-mediated gait training accompanied with or without virtual reality program. Objectives: Exploring whether the results indicated virtual reality system has contribution to muscle strength and balance ability. Design: A case series research, cross-over trial. Methods: Eleven participants (male 4, female 7) with adults diagnosed as stroke from medical doctor ware engaged. The participants received 2 treatment sessions of identical duration, robot-assisted gait training with virtual reality and robot-assisted gait training with screen-off randomly crossed over include 1-day for each person of wash-out period. The parameter was muscle activity, the researchers assessed sEMG (surface electromyography). Results: The result showed less muscle activities during training in robot-assisted gait training with virtual reality circumstances, and these indicated muscles were gluteus medius muscle, vastus medialis muscle, vastus intermedius and vastus lateralis muscle, semimembranosus muscle, gastrocnemius-lateral head, and soleus muscle (P<.05). Conclusion: In this study, we analyzed the outcome of muscle activity for clinical inference of robot-assisted gait training with virtual reality (VR). Less muscle activity was measured in the treatment accompanied by VR, therefore, a more systematic, in-depth and well-founded level of follow-up research is needed.

• International Journal of Industrial Entomology and Biomaterials
• /
• v.11 no.2
• /
• pp.119-124
• /
• 2005

To determine the level of heterosis, higher cocoon shell weight multivoltine congenic lines (Con. L) and bivoltine syngenic lines (Syn. L) of silkworm were used for crosses. First filial generations $(F_1s)$ expressed heterobeltiotic genetic interaction at significant magnitude (p < 0.01) for single cocoon shell weight (SCSW). The other linked characters viz., single cocoon weight (SCW) and yield by weight per 10, 000 larvae were also significantly higher (p < 0.01) than the better parental lines. All the hybrids showed significant improvement for these aforesaid characters over standard heterosis (Standard check). The reeling parameters viz., filament length, raw silk, neatness, cohesionstrokes etc, also showed improvement among the hybrids than check in congenial environment. Overall results suggested that the cross between congenic and syngenic lines provide better heterosis with good quality silk than conventional hybrids and may be used for commercial exploitation.

• Nuclear Engineering and Technology
• /
• v.54 no.8
• /
• pp.3027-3033
• /
• 2022

Background: In this study, various types of deep-learning models for predicting in vitro radiosensitivity from gene-expression profiling were compared. Methods: The clonogenic surviving fractions at 2 Gy from previous publications and microarray gene-expression data from the National Cancer Institute-60 cell lines were used to measure the radiosensitivity. Seven different prediction models including three distinct multi-layered perceptrons (MLP), four different convolutional neural networks (CNN) were compared. Folded cross-validation was applied to train and evaluate model performance. The criteria for correct prediction were absolute error < 0.02 or relative error < 10%. The models were compared in terms of prediction accuracy, training time per epoch, training fluctuations, and required calculation resources. Results: The strength of MLP-based models was their fast initial convergence and short training time per epoch. They represented significantly different prediction accuracy depending on the model configuration. The CNN-based models showed relatively high prediction accuracy, low training fluctuations, and a relatively small increase in the memory requirement as the model deepens. Conclusion: Our findings suggest that a CNN-based model with moderate depth would be appropriate when the prediction accuracy is important, and a shallow MLP-based model can be recommended when either the training resources or time are limited.

• Physical Therapy Rehabilitation Science
• /
• v.11 no.4
• /
• pp.571-578
• /
• 2022

Objective: Cognitive-motor trainings had a positive impact on cognitive function and dual-task trainings led to improvements of global cognitive function. The brain activity of the prefrontal cortex (PFC) is another indicator that can infer cognitive function. This study aims to confirm the usability of the interactive system cognitive-motor training program and the changes in the prefrontal cortex through training. Design: Cross-sectional study Methods: In this study, two cognitive tasks were randomly applied to 20 adults as cognitive-motor training using an interactive system, and the same task was performed using the original method. During all tasks, the brain activity of the prefrontal cortex was measured by the change in oxyhemoglobin (HbO) in real-time using Functional Near-Infrastructure. After performing the tasks, the usability of the developed interactive system was evaluated by a usability questionnaire which consists of five items, and each item consists of a 7-point Likert scale that responds from 1 point to 7 points. Results: The HbO levels were increased during cognitive task performance than at the resting phase. And evaluating the usefulness of the interactive system, a questionnaire result showed that it would be useful for cognitive-motor programs. Conclusions: The cognitive-motor training using the interactive system increased the activity of the prefrontal cortex, and the developed wearable sensor-based interactive system confirmed its usefulness.

• Journal of the Korean Society of Physical Medicine
• /
• v.12 no.1
• /
• pp.9-14
• /
• 2017

PURPOSE: To evaluate the effect of weight shift training with Hula Hoop on weight shift change and gait in stroke patients. METHODS: Ten stroke patients were enrolled in this study, and randomly divided into 2 groups. The study group underwent weight shift training with Hula Hoop, while the control group received general physical therapy that included weight shift training. All the studies were performed over a period of 4 weeks. Before and after the intervention, plantar pressure and performance in the 10 m walk test (10MWT) were assessed. Wilcoxon signed ranks test was used to compare the change from before to after the intervention in each group. The differences between the study and control groups were analyzed by using the Mann-Whitney test. RESULTS: After 4 weeks of intervention, the change in weight shift and performance in the 10MWT from before to after the experiment showed no statistical significance (p>.05). In addition, the comparison between the groups showed no significance in terms of weight-shift change, and performance in the 10MWT (p>.05). CONCLUSION: Although the difference was not statistically significant, the degree of improvement was similar to that attained with the conventional exercise treatment related to weight- shift training. During the course of the treatment, the patients received feedback through repeated training by themselves. Weight-shift training with Hula Hoop would be effective in improving the walking ability and weight-shifting on the paralyzed side of stroke patients. In the future, the effectiveness of this training would need to be validated.

• Asian Pacific Journal of Cancer Prevention
• /
• v.15 no.14
• /
• pp.5523-5528
• /
• 2014

Objective: To evaluate the implementation of screening colonoscopy amongst first-degree relatives (FDRs) of patients with colorectal cancer (CRC) in Turkey. Materials and Methods: A total of 400 first-degree relatives (mean(SD)age: 42.5(12.7) years, 55.5% were male) of 136 CRC patients were included in this cross-sectional questionnaire based survey. Data on demographic characteristics, relationship to patient and family history for malignancy other than the index case were evaluated in the FDRs of patients as were the data on knowledge about and characteristics related to the implementation of screening colonoscopy using a standardized questionnaire form. Results: The mean(SD) age at diagnosis of CRC in the index patients was 60.0(14.0) years, while mean(SD) age of first degree relatives was 42.5(12.7) years. Overall 36.3% of relatives were determined to have knowledge about colonoscopy. Physicians (66.9%) were the major source of information. Screening colonoscopy was recommended to 19.5% (n=78) of patient relatives, while 48.7% (n=38) of individuals participated in colonoscopy procedures, mostly (57.9%) one year after the index diagnosis. Screening colonoscopy revealed normal findings in 25 of 38 (65.8%) cases, while precancerous lesions were detected in 26.3% of screened individuals. In 19.0% of FDRs of patients, there was a detected risk for Lynch syndrome related cancer. Conclusions: In conclusion, our findings revealed that less than 20% of FDRs of patients had received a screening colonoscopy recommendation; only 48.7% participated in the procedure with detection of precancerous lesions in 26.3%. Rise of awareness about screening colonoscopy amongst patients with CRC and first degree relatives of patients and motivation of physicians for targeted screening would improve the participation rate in screening colonoscopy by FDRs of patients with CRC in Turkey.

• International Journal of Internet, Broadcasting and Communication
• /
• v.16 no.1
• /
• pp.148-155
• /
• 2024

We are designed was to find an efficient warm-up method for resistance training for muscle hypertrophy, and 10 males with at least 3 years of resistance training experience were selected as subjects. The 75% 1RM was measured directly based on the pre-measured bench press 1RM. After that, the main experiment of 75% 1RM bench press according to the three warm-up methods was conducted one week apart, and all experiments were randomized and cross-over. Performance according to the warm-up method (3) was measured by total exercise volume, and physiological changes were determined by myocardial workload. All post-measurement data were analyzed using SPSS.22.0 and analyzed using repeated measures one-way ANOVA and contrast comparisons were made using the deviation method. The results showed that the method of gradually increasing the number of repetitions by performing the same intensity as the intensity of the main exercise in the form of muscle hypertrophy, but at submaximal repetitions, showed the highest performance.

• International Journal of Contents
• /
• v.11 no.2
• /
• pp.57-62
• /
• 2015

Relative entropy is a divergence measure between two probability density functions of a random variable. Assuming that the random variable has only two alphabets, the relative entropy becomes a cross-entropy error function that can accelerate training convergence of multi-layer perceptron neural networks. Also, the n-th order extension of cross-entropy (nCE) error function exhibits an improved performance in viewpoints of learning convergence and generalization capability. In this paper, we derive a new divergence measure between two probability density functions from the nCE error function. And the new divergence measure is compared with the relative entropy through the use of three-dimensional plots.

• Journal of the Korean Data and Information Science Society
• /
• v.27 no.2
• /
• pp.549-557
• /
• 2016

In this paper we propose multiclassification method for large data sets by ensembling least squares support vector machines (LS-SVM) with principal components instead of raw input vector. We use the revised one-vs-all method for multiclassification, which is one of voting scheme based on combining several binary classifications. The revised one-vs-all method is performed by using the hat matrix of LS-SVM ensemble, which is obtained by ensembling LS-SVMs trained using each random sample from the whole large training data. The leave-one-out cross validation (CV) function is used for the optimal values of hyper-parameters which affect the performance of multiclass LS-SVM ensemble. We present the generalized cross validation function to reduce computational burden of leave-one-out CV functions. Experimental results from real data sets are then obtained to illustrate the performance of the proposed multiclass LS-SVM ensemble.