• Speech Sciences
• /
• v.11 no.1
• /
• pp.7-20
• /
• 2004

In this paper, we propose a two-step noise compensation algorithm in feature extraction for achieving robust speech recognition. The proposed method frees us from requiring a priori information on noisy environments and is simple to implement. First, in frequency domain, the Harmonics-based Spectral Subtraction (HSS) is applied so that it reduces the additive background noise and makes the shape of harmonics in speech spectrum more pronounced. We then apply a judiciously weighted variance Feature Vector Normalization (FVN) to compensate for both the channel distortion and additive noise. The weighted variance FVN compensates for the variance mismatch in both the speech and the non-speech regions respectively. Representative performance evaluation using Aurora 2 database shows that the proposed method yields 27.18% relative improvement in accuracy under a multi-noise training task and 57.94% relative improvement under a clean training task.

• Proceedings of the Korean Magnestics Society Conference
• /
• 2007.05a
• /
• pp.160.1-160.1
• /
• 2007

• Journal of Industrial Convergence
• /
• v.19 no.4
• /
• pp.111-126
• /
• 2021

The purpose of this study was to investigate the effects of 12-week training on changes in physical fitness and cardiovascular factors for firefighters. For this purpose, 40 men in their 20s and 30s who agreed to participate voluntarily were recruited. They were divided into four groups: the firefighters' physical fitness test training group (hereinafter referred to as PT group), firefighters' physical fitness test and aerobic training group (hereinafter referred to as PT+AR group), firefighters' physical fitness test and both aerobic and anaerobic training group (hereinafter referred to as PT+CO group). Physical fitness factors (grip strength, back muscle strength, seated forward bend, standing long jump, sit-ups, 20-meter shuttle run), cardiovascular factors (total cholesterol, triglycerides, high-density lipoprotein cholesterol, low-density lipoprotein cholesterol, glucose, waist circumference, systolic blood pressure, diastolic blood pressure) and the relationship between Framingham Heart Risk Score and physical/cardiovascular factors were compared and analyzed, and the following conclusions were obtained. Aerobic training, anaerobic training, and combined training, including 12 weeks of firefighter physical examinations, all had positive effects on fitness and cardiovascular factors, which would be an appropriate way for firefighter examinees to improve physical strength and reduce the risk of cardiovascular disease.

• Proceedings of the Korean Magnestics Society Conference
• /
• 2009.12a
• /
• pp.98-99
• /
• 2009

• Journal of Physiology & Pathology in Korean Medicine
• /
• v.21 no.5
• /
• pp.1356-1363
• /
• 2007

The purpose of the study was to identify the general status of Kigong organizations introduced on Internet. We have used www.naver.com, the biggest portal site in Korea and www.nice114.co.kr, which has the longest list of the telephone numbers to look up the organizations with the index of "Kigong and Danhak" and "Mediation". Among them we screened the organizations to have the list of organizations which have been established for more than 5 years, with more than 100 trainees and whether they published books or booklets regarding Kigong by the means of telephone conversation or home page access. The number of organizations identified by telephone and Internet with the indexes of Kigongdanhak and mediation was 852. The number of organizations that passed the primary criterion was 22, and that passed the secondary criterion was 8. Among the primarily screened organization, there are 5 focusing on mediation, 5 focusing on breathing, 3 focusing on Haenggong, 4 focusing on mediation and Haenggong, 4 focusing on breathing and Haenggong and 1 focusing on mediation and breathing. In secondarily screened organizations, they called their training method as Seondo, Shinseondo or Seonhak and origin of the training method as Dangun and Hwangwung. As Sambeop training of Jigam, Josik and Geumchok provide training methods which are a little different each other, the utilization rate was low although there are some organizations that have special training using Three Bibles. It was identified that there were many texts and writings that they took as training methods other than Three Bibles. Kigong training organizations based on Korean traditional ideologies are grounded on the Three Textbooks such as , , and and the concept of Hongikingan. This ideological ground is the study of Completion of Human Beings through physical and mental training and goes with Seondo, Pungryudo and Hyunmyojido.

• Journal of Engineering Education Research
• /
• v.11 no.2
• /
• pp.65-78
• /
• 2008

Currently co-work programs between industries and engineering schools play an important role for mutual interests, win-win strategy. Industries can develop new technologies through the human resources and facilities in school and schools are able to have research achievements by applying their own theoretical abilities to real field of industries's projects. However most co-work programs between industries and engineering schools mainly focus on programs such as research projects with graduate research lab, on-site training for job and training program for field engineers. And more it is difficult for schools to make co-work programs targeted engineering education for undergraduate school students because of many constraints such as planing program, budget and indifference of companies. Therefore this paper introduces LG Electronics Display Idea Competition hosted by LG Display Division in Kumi, S. Korea and present what benefits to both school and company made through this program and unique model of co-work program for engineering education between school and company in country.

• Journal of IKEEE
• /
• v.26 no.4
• /
• pp.531-536
• /
• 2022

In order to efficeintly manage a battery, it is important to accurately estimate and manage the SOH(State of Health) and RUL(Remaining Useful Life) of the batteries. Even if the batteries are of the same type, the characteristics such as facility capacity and voltage are different, and when the battery for the training model and the battery for prediction through the model are different, there is a limit to measuring the accuracy. In this paper, We proposed the entropy index using voltage distribution and discharge time is generalized, and four batteries are defined as a training set and a test set alternately one by one to predict the health status of batteries through linear regression analysis of machine learning. The proposed method showed a high accuracy of more than 95% using the MAPE(Mean Absolute Percentage Error).

• Journal of information and communication convergence engineering
• /
• v.22 no.2
• /
• pp.165-171
• /
• 2024

In this study, we present a novel approach for enhancing chest X-ray image classification (normal, Covid-19, edema, mass nodules, and pneumothorax) by combining contrastive learning and machine learning algorithms. A vast amount of unlabeled data was leveraged to learn representations so that data efficiency is improved as a means of addressing the limited availability of labeled data in X-ray images. Our approach involves training classification algorithms using the extracted features from a linear fine-tuned Momentum Contrast (MoCo) model. The MoCo architecture with a Resnet34, Resnet50, or Resnet101 backbone is trained to learn features from unlabeled data. Instead of only fine-tuning the linear classifier layer on the MoCopretrained model, we propose training nonlinear classifiers as substitutes for softmax in deep networks. The empirical results show that while the linear fine-tuned ImageNet-pretrained models achieved the highest accuracy of only 82.9% and the linear fine-tuned MoCo-pretrained models an increased highest accuracy of 84.8%, our proposed method offered a significant improvement and achieved the highest accuracy of 87.9%.

• Korean Journal of Remote Sensing
• /
• v.39 no.5_1
• /
• pp.621-635
• /
• 2023

The assessment of structural condition is a crucial process for evaluating its usability and determining the diagnostic cycle. The currently employed manpower-based methods suffer from issues related to safety, efficiency, and objectivity. To address these concerns, research based on deep learning using images is being conducted. However, acquiring structural damage data is challenging, making it difficult to construct a substantial amount of training data, thus limiting the effectiveness of deep learning-based condition assessment. In this study, we propose a foundation model-based 2-step structural damage analysis to overcome the lack of training data in image-based structural condition assessments. We subdivided the elements of structural condition assessment into instantiation and quantification. In the quantification step, we applied a foundation model for image segmentation. Our method demonstrated a 10%-point increase in mean intersection over union compared to conventional image segmentation techniques, with a notable 40%-point improvement in the case of rebar exposure. We anticipate that our proposed approach will enhance performance in domains where acquiring training data is challenging.

• Journal of Power System Engineering
• /
• v.20 no.4
• /
• pp.19-25
• /
• 2016

It is needed to use the gait training system for the rehabilitation of the disabled and old people. In this study, a gait training system of turn roller type is proposed for the purpose of helping the rehabilitation. A driving mechanism with the turn roller is designed by using the RecurDyn which is the dynamic analysis program. RecurDyn is used to analyze the dynamic behavior of the gait training system. The static load analysis is carried out to investigate the safety of this system. From the operating test of this system, it is noted that the driving error is little and the load capacity is 130 kgf.