• Journal of Navigation and Port Research
• /
• v.27 no.5
• /
• pp.499-504
• /
• 2003

This paper describes the system for evaluating the sea wave informations such as wave direction and wave length in real time, by using image data obtained from the marine X-band radar. We proposed here a method for automatic selection of the partial image data without the user's individual selection at the radar. We also discussed that the wave direction could be obtained by a 2-dimensional discrete Fourier transform algorithm. We carried some evaluation works on the algorithm through computer simulation. The obtained thirteen radar image data under several sea surface conditions were analyzed by the method described and the result was presented.

• Progress in Medical Physics
• /
• v.14 no.3
• /
• pp.203-210
• /
• 2003

PET (positron emission tomography) permits the investigation of physiological and biochemical processes in vivo. The accuracy of quantifying PET data is affected by its finite spatial resolution, which causes partial volume effects. In this study, we developed a method for partial volume correction using Hoffman phantom PET and MR data, and applied various FWHM (full width at half maximum) levels. We also applied this method to PET images of normal controls and tested for the possibility of clinical application. $^{18}$ F-PET Hoffman phantom images were co-registered to MR slices. The gray matter and white matter regions were then segmented into binary images. Each binary image was convolved by 4, 8, 12, 16 mm FWHM levels. These convolved images of gray and white matter were merged corresponding to the same level of FWHM. The original PET images were then divided by the convolved binary images voxel-by-voxel. These corrected PET images were multiplied by binary images. The corrected PET images were evaluated by analyzing regions of interests, which were drawn on the gray and white matter regions of the original MR image slices. We calculated the ratio of white to gray matter. We also applied this method to the PET images of normal controls. On analyzing the corrected PET images of Hoffman phantom, the ratios of the corrected images increased more than that of the uncorrected images. With the normal controls, the ratio of the corrected images increased more than that of the uncorrected images. The ratio increase of the corrected PET images was lower than that of the corrected phantom PET images. In conclusion, the method developed for partial volume correction in PET data may be clinically applied, although further study may be required for optimal correction.

• Journal of the Korea Society of Computer and Information
• /
• v.16 no.1
• /
• pp.47-52
• /
• 2011

This paper collect sleep environment data of bedroom to sleeping, and analyzing the relationship between conditions with obtained data and sleep. We provide the optimal sleep environment of individual by extracting the simulation model based on it. The experiments was using temperature/humidity sensor(SHT11) and ambient light sensors(GL5507). For extraction of tossing and turning, we use difference image method in motion extraction from video. In addition, the information of weight can affect to sleep, it was entered such as ratio of fatigue, drinking, empty stomach. As a result, we are able to extract the optimal sleep environment. The future, we will try to improve to help to lead more pleasant daily life providing proper indoor environment changes depending on the situation even a partial of organic ubiquitous living environments such as eating, work ete. as well as certain sleep circumstances.

• Journal of Korean Association for Spatial Structures
• /
• v.21 no.2
• /
• pp.99-110
• /
• 2021

This study presents the estimation of crack depth by analyzing temperatures extracted from thermal images and environmental parameters such as air temperature, air humidity, illumination. The statistics of all acquired features and the correlation coefficient among thermal images and environmental parameters are presented. The concrete crack depths were predicted by four different machine learning models: Multi-Layer Perceptron (MLP), Random Forest (RF), Gradient Boosting (GB), and AdaBoost (AB). The machine learning algorithms are validated by the coefficient of determination, accuracy, and Mean Absolute Percentage Error (MAPE). The AB model had a great performance among the four models due to the non-linearity of features and weak learner aggregation with weights on misclassified data. The maximum depth 11 of the base estimator in the AB model is efficient with high performance with 97.6% of accuracy and 0.07% of MAPE. Feature importances, permutation importance, and partial dependence are analyzed in the AB model. The results show that the marginal effect of air humidity, crack depth, and crack temperature in order is higher than that of the others.

• School Mathematics
• /
• v.5 no.2
• /
• pp.259-273
• /
• 2003

A fraction is one of the most important concepts that students have to learn in elementary school. But it is a challenge for students to understand fraction concept because of its conceptual complexity. The focus of fraction learning is understanding the concept. Then the problem is how we can facilitate the conceptual understanding and estimate it. In this study, Moore's concept understanding scheme(concept definition, concept image, concept usage) was adopted as an theoretical framework to investigate students' fraction understanding. The questions of this study were a) what concept image do students have\ulcorner b) How well do students solve fraction problems\ulcorner c) How do students use fraction concept to generate fraction word problem\ulcorner By analyzing the data gathered from three elementary school, several conclusion was drawn. 1) The students' concept image of fraction is restricted to part-whole sub-construct. So is students' fraction understanding. 2) Students can solve part-whole fraction problems well but others less. This also imply that students' fraction understanding is partial. 3) Half of the subject(N=98) cannot pose problems that involve fraction and fraction operation. And some succeeded applied the concept mistakenly. To understand fraction, various fraction subconstructs have to be integrated as whole one. To facilitate this integration, fraction program should focus on unit, partitioning and quantity. This may be achieved by following activities: * Building on informal knowledge of fraction * Focusing on meaning other than symbol * Various partitioning activities * Facing various representation * Emphasizing quantitative aspects of fraction * Understanding the meanings of fraction operation Through these activities, teacher must help students construct various faction concept image and apply it to meaningful situation. Especially, to help students to construct various concept image and to use fraction meaningfully to pose problems, much time should be spent to problem posing using fraction.