• Journal of the Korean Geographical Society
• /
• v.51 no.2
• /
• pp.219-233
• /
• 2016

In social science fields, statistical models are used almost exclusively for causal explanation, and explanatory modeling has been a mainstream until now. In contrast, predictive modeling has been rare in the fields. Hence, we focus on constructing the predictive non-parametric model, instead of the explanatory model. Gangnam-gu, Seoul was chosen as a study area and we collected single-family house sales data sold between 2011 and 2014. We applied non-parametric models proposed in machine learning area including generalized additive model(GAM), random forest, multivariate adaptive regression splines(MARS) and support vector machines(SVM). Models developed recently such as MARS and SVM were found to be superior in predictive power for house price estimation. Finally, spatial autocorrelation was accounted for in the non-parametric models additionally, and the result showed that their predictive power was enhanced further. We hope that this study will prompt methodology for property price estimation to be extended from traditional parametric models into non-parametric ones.

• Structural Engineering and Mechanics
• /
• v.83 no.3
• /
• pp.327-340
• /
• 2022

The Collapse Margin Ratio (CMR) is a notable index used for seismic assessment of the structures. As proposed by FEMA P695, a set of analyses including the Nonlinear Static Analysis (NSA), Incremental Dynamic Analysis (IDA), together with Fragility Analysis, which are typically time-taking and computationally unaffordable, need to be conducted, so that the CMR could be obtained. To address this issue and to achieve a quick and efficient method to estimate the CMR, the Artificial Neural Network (ANN), Response Surface Method (RSM), and Adaptive Neuro-Fuzzy Inference System (ANFIS) will be introduced in the current research. Accordingly, using the NSA results, an attempt was made to find a fast and efficient approach to derive the CMR. To this end, 5016 IDA analyses based on FEMA P695 methodology on 114 various Reinforced Concrete (RC) frames with 1 to 12 stories have been carried out. In this respect, five parameters have been used as the independent and desired inputs of the systems. On the other hand, the CMR is regarded as the output of the systems. Accordingly, a double hidden layer neural network with Levenberg-Marquardt training and learning algorithm was taken into account. Moreover, in the RSM approach, the quadratic system incorporating 20 parameters was implemented. Correspondingly, the Analysis of Variance (ANOVA) has been employed to discuss the results taken from the developed model. Additionally, the essential parameters and interactions are extracted, and input parameters are sorted according to their importance. Moreover, the ANFIS using Takagi-Sugeno fuzzy system was employed. Finally, all methods were compared, and the effective parameters and associated relationships were extracted. In contrast to the other approaches, the ANFIS provided the best efficiency and high accuracy with the minimum desired errors. Comparatively, it was obtained that the ANN method is more effective than the RSM and has a higher regression coefficient and lower statistical errors.

• Journal of Information Processing Systems
• /
• v.18 no.6
• /
• pp.729-740
• /
• 2022

Netflix, Amazon Prime, and YouTube are the most popular and fastest-growing streaming services globally. It is a matter of great interest for the streaming service providers to preview their service infrastructure and streaming strategy in order to provide new streaming services. Hence, the first part of the paper presents a detailed survey of the Content Distribution Network (CDN) and cloud infrastructure of these service providers. To understand the streaming strategy of these service providers, the second part of the paper deduces a common quality-of-service (QoS) model based on rebuffering time, bitrate, progressive download ratio, and standard deviation of the On-Off cycle. This model is then used to analyze and compare the streaming behaviors of these services. This study concluded that the streaming behaviors of all these services are similar as they all use Dynamic Adaptive Streaming over HTTP (DASH) on top of TCP. However, the amount of data that they download in the buffering state and steady-state vary, resulting in different progressive download ratios, rebuffering levels, and bitrates. The characteristics of their On-Off cycle are also different resulting in different QoS. Hence a thorough adaptive bit rate (ABR) analysis is presented in this paper. The streaming behaviors of these services are tested on different access network bandwidths, ranging from 75 kbps to 30 Mbps. The survey results indicate that Netflix QoS and streaming behavior are significantly consistent followed by Amazon Prime and YouTube. Our approach can be used to compare and contrast the streaming services' strategies and finetune their ABR and flow control mechanisms.

• KSCE Journal of Civil and Environmental Engineering Research
• /
• v.26 no.4A
• /
• pp.677-687
• /
• 2006

Kriging interpolation is one of the generally used interpolation techniques in Geostatistics field. This technique includes the experimental and theoretical variograms and the formulation of kriging interpolation. In contrast to the conventional least square method for stress recovery, kriging interpolation is based on the weighted least square method to obtain the estimated exact solution from the stress data at the Gauss points. The weight factor is determined by variogram modeling for interpolation of stress data apart from the conventional interpolation methods that use an equal weight factor. In addition to this, the p-level is increased non-uniformly or selectively through a posteriori error estimation based on SPR (superconvergent patch recovery) technique, proposed by Zienkiewicz and Zhu, by auto mesh p-refinement. The cut-out plate problem under tension has been tested to validate this approach. It also provides validity of kriging interpolation through comparing to existing least square method.

• Journal of the Korea Society of Computer and Information
• /
• v.29 no.2
• /
• pp.101-108
• /
• 2024

AI-OCR (Artificial Intelligence Optical Character Recognition) combines OCR technology with Artificial Intelligence to overcome limitations that required human intervention. To enhance the performance of AI-OCR, training on diverse data sets is essential. However, the recognition rate declines when image colors have similar brightness levels. To solve this issue, this study employs Homomorphic filtering as a preprocessing step to clearly differentiate color levels, thereby increasing text recognition rates. While Homomorphic filtering is ideal for text extraction because of its ability to adjust the high and low frequency components of an image separately using a gamma value, it has the downside of requiring manual adjustments to the gamma value. This research proposes a range for gamma threshold values based on tests involving image contrast, brightness, and entropy. Experimental results using the proposed range of gamma values in Homomorphic filtering suggest a high likelihood for effective AI-OCR performance.

• Journal of International Society for Simulation Surgery
• /
• v.1 no.1
• /
• pp.37-40
• /
• 2014

Purpose For living donor liver transplantation, liver segmentation is difficult due to the variability of its shape across patients and similarity of the density of neighbor organs such as heart, stomach, kidney, and spleen. In this paper, we propose an automatic segmentation of the liver using multi-planar anatomy and deformable surface model in portal phase of abdominal contrast-enhanced CT images. Method Our method is composed of four main steps. First, the optimal liver volume is extracted by positional information of pelvis and rib and by separating lungs and heart from CT images. Second, anisotropic diffusing filtering and adaptive thresholding are used to segment the initial liver volume. Third, morphological opening and connected component labeling are applied to multiple planes for removing neighbor organs. Finally, deformable surface model and probability summation map are performed to refine a posterior liver surface and missing left robe in previous step. Results All experimental datasets were acquired on ten living donors using a SIEMENS CT system. Each image had a matrix size of $512{\times}512$ pixels with in-plane resolutions ranging from 0.54 to 0.70 mm. The slice spacing was 2.0 mm and the number of images per scan ranged from 136 to 229. For accuracy evaluation, the average symmetric surface distance (ASD) and the volume overlap error (VE) between automatic segmentation and manual segmentation by two radiologists are calculated. The ASD was $0.26{\pm}0.12mm$ for manual1 versus automatic and $0.24{\pm}0.09mm$ for manual2 versus automatic while that of inter-radiologists was $0.23{\pm}0.05mm$. The VE was $0.86{\pm}0.45%$ for manual1 versus automatic and $0.73{\pm}0.33%$ for manaual2 versus automatic while that of inter-radiologist was $0.76{\pm}0.21%$. Conclusion Our method can be used for the liver volumetry for the pre-surgery planning of living donor liver transplantation.

• Journal of the Institute of Electronics Engineers of Korea SP
• /
• v.37 no.1
• /
• pp.37-48
• /
• 2000

In this paper, we propose a new connected operator Using morphological grayscale reconstruction for region-based coding First, an effective method of reference-image creation lis proposed, which is based on the Size as well as the contrast. This improves the performance of simplification, because It preserves perceptually important components and removes unnecessary components The conventional connected operators are good for removing small regions, but have a serious drawback for low-contrast regions that are larger than the structuring element. That is, when the conventional connected operators are applied to tills region, the simplification becomes less effective or several meaningful regions are merged to one region to avoid this, the conventional geodesic dilation is modified to propose an adaptive operator to reduce the effect of inappropriate propagation, pixels reconstructed to the original values are excluded m the dilation operation Experimental results have shown that the proposed algorithm achieves better performance In terms of the reconstruction of flat zones. The Picture quality has also been improved by about 7dB, compared to the conventional methods.

• Journal of radiological science and technology
• /
• v.47 no.1
• /
• pp.29-37
• /
• 2024

This study aims to evaluate the performance of the U-Net based learning model that may vary depending on the histogram equalization algorithm. The subject of the experiment were 17 radiology students of this college, and 1,727 data sets in which the region of interest was set in the thyroid after acquiring ultrasound image data were used. The training set consisted of 1,383 images, the validation set consisted of 172 and the test data set consisted of 172. The equalization algorithm was divided into Histogram Equalization(HE) and Contrast Limited Adaptive Histogram Equalization(CLAHE), and according to the clip limit, it was divided into CLAHE8-1, CLAHE8-2. CLAHE8-3. Deep Learning was learned through size control, histogram equalization, Z-score normalization, and data augmentation. As a result of the experiment, the Attention U-Net showed the highest performance from CLAHE8-2 to 0.8355, and the U-Net and BSU-Net showed the highest performance from CLAHE8-3 to 0.8303 and 0.8277. In the case of mIoU, the Attention U-Net was 0.7175 in CLAHE8-2, the U-Net was 0.7098 and the BSU-Net was 0.7060 in CLAHE8-3. This study attempted to confirm the effects of U-Net, Attention U-Net, and BSU-Net models when histogram equalization is performed on ultrasound images. The increase in Clip Limit can be expected to increase the ROI match with the prediction mask by clarifying the boundaries, which affects the improvement of the contrast of the thyroid area in deep learning model learning, and consequently affects the performance improvement.

• Journal of radiological science and technology
• /
• v.40 no.2
• /
• pp.261-267
• /
• 2017

The study has attempted to evaluate and compare the image evaluation and exposure dose by respectively applying filter back projection (FBP), the existing test method, and adaptive statistical iterative reconstruction (ASIR) with different values of tube voltage during the low dose computed tomography (LDCT). With the image reconstruction method as basis, chest phantom was utilized with the FBP and ASIR set at 10%, 20% respectively, and the change of tube voltage (100 kVp, 120 kVp). For image evaluation, back ground noise, signal-noise ratio (SNR) and contrast-noise ratio (CNR) were measured, and, for dose assessment, CTDIvol and DLP were measured respectively. In terms of image evaluation, there was significant difference in ascending aorta (AA) SNR and inpraspinatus muscle (IM) SNR with the different amount of tube voltage (p < 0.05). In terms of CTDIvol, the measured values with the same tube voltage of 120 kVp were 2.6 mGy with no-ASIR and 2.17 mGy with 20%-ASIR respectively, decreased by 0.43 mGy, and the values with 100 kVp were 1.61 mGy with no-ASIR and 1.34 mGy with 20%-ASIR, decreased by 0.27 mGy. In terms of DLP, the measured values with 120 kVp were $103.21mGy{\cdot}cm$ with no-ASIR and $85.94mGy{\cdot}cm$ with 20%-ASIR, decreased by $17.27mGy{\cdot}cm$ (about 16.7%), and the values with 100 kVp were $63.84mGy{\cdot}cm$ with no-ASIR and $53.25mGy{\cdot}cm$ with 20%-ASIR, a decrease by $10.62mGy{\cdot}cm$ (about 16.7%). At lower tube voltage, the rate of dose significantly decreased, but the negative effects on image evaluation was shown due to the increase of noise.

• Radiation Oncology Journal
• /
• v.17 no.4
• /
• pp.307-313
• /
• 1999

Purpose : To investigate the Presence of adaptive response by low dose radiation in murine tumors in relation to radiation induced apoptosis as well as related mechanism. Materials and Methods : Syngeneic murine tumors, OCa-I and HCa-l, were given 0.05 Gy pretreatment followed by therapeutic dose of 25 Gy radiation. Induction of apoptosis was analyzed for each treatment group. Regulating molecules of apoptosis, p53, Bcl-2, Bax, Bcl-X, were also analyzed by Western blotting. Results : In 0.05 Gy pretreatment group of OCa-I, 25 Gy-induced apoptosis per 1000 cells was 229, which was estimated at $30\%$ lower level than the expected (p<0.05). In contrast, this reduction in radiation induced apoptosis was not seen in HCa-l. In the expression of apoptosis regulating molecules, p53 increased in both tumors in response to radiation. Bcl-2 and Bax did not show significant change in both tumors however, the expression of Bcl-2 surpassed that of Bax in 0.05 Gy pretreatment group of OCa-l. Bcl-X was not expressed in OCa-l. In HCa-l, Bcl-X showed increased expression even with 0.05 Gy. Conclusion : Adaptive response by low dose radiation Is shown in one murine tumor, OCa-l, in relation to radiation induced apoptosis. Apoptosis regulating molecules including Bcl-2/Bax and Bcl-X, appear to related. This study shows an evidence that adaptive response is present, but not a generalized phenomenon in vivo.