• Journal of IKEEE
• /
• v.22 no.4
• /
• pp.948-952
• /
• 2018

Haze removal is an object of scientific desire due to its various practical applications. Existing algorithms are founded upon histogram equalization, contrast maximization, or the growing trend of applying machine learning in image processing. Since machine learning-based algorithms solve problems based on the data, they usually perform better than those based on traditional image processing/computer vision techniques. However, to achieve such a high performance, one of the requisites is a large and reliable training database, which seems to be unattainable owing to the complexity of real hazy and haze-free images acquisition. As a result, researchers are currently using the synthetic database, obtained by introducing the synthetic haze drawn from the standard uniform distribution into the clear images. In this paper, we propose the enhanced equidistribution, improving upon our previous study on equidistribution, and use it to make a new database for training machine learning-based haze removal algorithms. A large number of experiments verify the effectiveness of our proposed methodology.

• ETRI Journal
• /
• v.44 no.2
• /
• pp.327-338
• /
• 2022

Classification models for human action recognition require robust features and large training sets for good generalization. However, data augmentation methods are employed for imbalanced training sets to achieve higher accuracy. These samples generated using data augmentation only reflect existing samples within the training set, their feature representations are less diverse and hence, contribute to less precise classification. This paper presents new data augmentation and action representation approaches to grow training sets. The proposed approach is based on two fundamental concepts: virtual video generation for augmentation and representation of the action videos through robust features. Virtual videos are generated from the motion history templates of action videos, which are convolved using a convolutional neural network, to generate deep features. Furthermore, by observing an objective function of the genetic algorithm, the spatiotemporal features of different samples are combined, to generate the representations of the virtual videos and then classified through an extreme learning machine classifier on MuHAVi-Uncut, iXMAS, and IAVID-1 datasets.

• Journal of Fisheries and Marine Sciences Education
• /
• v.27 no.2
• /
• pp.441-451
• /
• 2015

The one-year onboard training, which has been implemented since 1988 according to the STCW Convention in 1978, is a crucial part of cadets' training and education. The choice of training shipping company is important that like a internship. This basic study is for the satisfaction at the shipping companies that trains cadets the Mokpo National Maritime University. The satisfaction for onboard training consists of a conscious, educational and environmental parts with six demographic characteristics and carried out the questionnaire survey of 206 cadets in shipping companies. The low parts of satisfaction are implemented cross tabulation analysis and in-depth interview for realize the reason. In this paper, the satisfaction from environmental part($3.83{\pm}0.75$) is high but conscious part($3.35{\pm}0.86$) and educational part($3.33{\pm}0.79$) are low. In demographic characteristics, male in gender, apprentice officer in duty, others in kind of crew, container in kind of vessel and 30,000~100,000 tons in size of vessel are high satisfaction in general. Specifically, feeding service($4.02{\pm}0.75$) in environment, influence on future vision($3.65{\pm}0.84$) in consciousness, goal for training($3.50{\pm}0.65$) in education showed the highest level in satisfaction.

• Journal of Korean Library and Information Science Society
• /
• v.42 no.1
• /
• pp.145-164
• /
• 2011

University libraries in Korea are facing a crisis because of decline of users and librarian reduction from internet development and digital media. To solve this problem, author referred to the literatures on change management of company. Because there are many accumulating knowledge and experiences about company change management. It is necessary to lead the sympathetic sense of crisis by all university librarians. The vision should be made by all librarians' agreement. But the vision practice is more important than developing. In order to practice the vision, it is more useful to share the leadership with general manager and influential librarians in Korea university libraries. Furthermore, systematic leadership training for librarians is needed than ever.

• CELLMED
• /
• v.5 no.4
• /
• pp.26.1-26.3
• /
• 2015

A new physiotherapy named Ciliary Muscle Training method and one successful case had been reported first for treatment of Posner-Schlossman Syndrome (PSS), also known as Glaucomatocyclitic Crisis. When a person observed a back and forth moving object, the ciliary muscle would involuntarily do sometimes tense and sometimes relaxant exercise which can relieve the tensive state of the ciliary muscle of patients suffering PSS disease. Through this period of continued treatment, this ciliary muscle training method can improve the health status of the ciliary muscle and finally achieve the purpose of curing PSS disease. In this successful case, the patient's feedback result was very good after receiving our suggested ciliary muscle training method as an attempt to treat PSS, ill right eye of this PSS patient showed normal appearance without any differences from the other normal eye, his vision test values of naked eye were both 4.9 under nearly 200 degrees of myopia. High intraocular pressure (IOP) was mitigated, measured IOP value was in the normal (≦ 21 mmHg) and comfortable state without PSS relapse until now. Therefore, ciliary muscle training method could be a potential clinical treatment approach to PSS disease.

• Journal of the Korean Society for Precision Engineering
• /
• v.22 no.8 s.173
• /
• pp.192-198
• /
• 2005

In this paper, we propose a new training system for the improvement of equilibrium sense using unstable platform. The equilibrium sense, which provides orientation with respect to gravity, is important to integrate the vision, somatosensory and vestibular function to maintain the equilibrium sense of the human body. In order to improve the equilibrium sense, we developed the software program such as a block game, pingpong game using Visual C++. These training system for the equilibrium sense consists of unstable platform, computer interface and software program. The unstable platform was a simple structure of elliptical-type which included tilt sensor, wireless RF module and the device of power supply. To evaluate the effect of balance training, we measured and evaluated the parameters as the moving time to the target, duration to maintain cursor in the target of screen and the error between sine curve and acquired data. As a results, the moving time to the target and duration to maintain cursor in the target was improved through the repeating training of equilibrium sense. It was concluded that this system was reliable in the evaluation of equilibrium sense. This system might be applied to clinical use as an effective balance training system.

• Journal of Korean Ophthalmic Optics Society
• /
• v.18 no.4
• /
• pp.489-493
• /
• 2013

Purpose: This study was to investigate peripheral vision and provide people with macular degeneration with a guideline for eccentric viewing training. Methods: 30 adult subjects with normal vision took part in this study. The lateral area of $20^{\circ}$ eccentricity from the fovea of right eye was only used to measure the physiological summation. The target was sinusoidal vertical gratings within a circular aperture. The critical points in spatial and temporal summation was found to compare each other for 0.7 cpd and 3.0 cpd, respectively. Results: Critical duration and contrast sensitivity for 0.7 cpd were 540 ms and 1.1, and 315 ms and 0.98 for 3.0 cpd respectively. The critical degrees and contrast sensitivity for 0.7 cpd were $11.3^{\circ}$ and 2.8, and $5^{\circ}$ and 2.63 for 3.0 cpd respectively. Conclusions: The critical point in peripheral vision reaches relatively faster than the one in central vision. It is recommended to train the peripheral retina under the lower spatial frequency more frequently for a short time than constantly for a long time.

• Journal of Korea Multimedia Society
• /
• v.20 no.5
• /
• pp.769-781
• /
• 2017

Deep learning networks like Convolutional Neural Networks (CNNs) show successful performances in many computer vision applications such as image classification, object detection, and so on. For implementation of deep learning networks in embedded system with limited processing power and memory, deep learning network may need to be simplified. However, simplified deep learning network cannot learn every possible scene. One realistic strategy for embedded deep learning network is to construct a simplified deep learning network model optimized for the scene images of the installation place. Then, automatic training will be necessitated for commercialization. In this paper, as an intermediate step toward automatic training under fisheye camera environments, we study more precise human localization in fisheye images, and propose an accurate human localization method, Automatic Ground-Truth Labelling Method (AGTLM). AGTLM first localizes candidate human object bounding boxes by utilizing GoogLeNet-LSTM approach, and after reassurance process by GoogLeNet-based CNN network, finally refines them more correctly and precisely(tightly) by applying saliency object detection technique. The performance improvement of the proposed human localization method, AGTLM with respect to accuracy and tightness is shown through several experiments.

• Proceedings of the Korean Society for Agricultural Machinery Conference
• /
• 2000.11c
• /
• pp.715-721
• /
• 2000

In this research, rule and neuro net based boundary extraction algorithm was developed. Extracting boundary of the interest, lean tissue, is essential for the quality evaluation of the beef based on color machine vision. Major quality features of the beef are size, marveling state of the lean tissue, color of the fat, and thickness of back fat. To evaluate the beef quality, extracting of loin parts from the sectional image of beef rib is crucial and the first step. Since its boundary is not clear and very difficult to trace, neural network model was developed to isolate loin parts from the entire image input. At the stage of training network, normalized color image data was used. Model reference of boundary was determined by binary feature extraction algorithm using R(red) channel. And 100 sub-images(selected from maximum extended boundary rectangle 11${\times}$11 masks) were used as training data set. Each mask has information on the curvature of boundary. The basic rule in boundary extraction is the adaptation of the known curvature of the boundary. The structured model reference and neural net based boundary extraction algorithm was developed and implemented to the beef image and results were analyzed.

• Journal of Korean Academy of Nursing
• /
• v.30 no.3
• /
• pp.647-658
• /
• 2000

The purpose of this study was to determine the effect of Korean traditional dance movement training on balance, gait and leg strength in elderly women who are forced to remain at home. Fifteen elderly women of an experimental group between the ages 65 and 75 years who have normal vision and passed the hearing and Romberg test, participated in the 12 weeks' dance movement training. Fourteen subjects of a control group were selected. Korean traditional dance movement training was developed on the basis of Korean traditional dance and music by the authors. It took approximately 50 minutes to perform the dance movement program. The subjects of the experimental group practiced dance training for 3 times a week during 12 weeks. During the 50 minutes workout, the subjects practiced 15 minutes of a warm-up dance, 25 minutes of a conditioning dance, and 10 minutes of a cool-down dance. The intensity for the conditioning phase was between 60% and 65% of age-adjusted maximum heart rates. The balance, gait and leg strength were measured prior to and after the experimental treatment. Total balance scores of the experimental group were significantly higher than those of the control group. Scores of sternal nudge, one leg standing balance and reaching up among 13 items have significantly increased after the dance movement training. Total scores of gait of the experimental group were significantly higher than those of the control group following the korean traditional dance movement training. Scores of experimental group in step height, path deviation and turning while walking among 9 items have increased significantly following 12 weeks of dance movement training. The leg strength of experimental group was significantly higher than those of the control group following the Korean traditional dance movement training. The balance, gait and leg strength have significantly correlated in the experimental group following the Korean traditional dance movement training. The results suggest that Korean traditional dance movement training can improve balance, gait and leg strength in home bound elderly women.