• Journal of the Earthquake Engineering Society of Korea
• /
• v.22 no.6
• /
• pp.323-332
• /
• 2018

Recent two moderate earthquakes (2016 $M_w=5.4$ Gyeongju and 2017 $M_w=5.5$ Pohang) in Korea provided the unique chance of developing a set of relations to estimate instrumental seismic intensity in Korea by augmenting the time-history data from MMI seismic intensity regions above V to the insufficient data previously accumulated from the MMI regions limited up to IV. The MMI intensity regions of V and VI was identified by delineating the epicentral distance from the reference intensity statistics in distance derived by using the integrated MMI data obtained by combining the intensity survey results of KMA (Korea Meteorological Administration) and 'DYFI (Did You Feel It)' MMIs of USGS. The time-histories of the seismic stations from the MMI intensity regions above V were then preprocessed by applying the previously developed site-correction filters to be converted to a site-equivalent condition in a manner consistent with the previous study. The average values of the ground-motion parameters for the three ground motion parameters of PGA, PGV and BSPGA (Bracketed Summation of PGA per second for 30 seconds) were calculated for the MMI=V and VI and used to generate the dataset of the average values of the ground-motion parameters for the individual MMIs from I to VI. Based on this dataset, the linear regression analysis resulted in the following relations with proposed valid ranges of MMI. $MMI=2.36{\times}log_{10}(PGA(gal))+1.44$ ($I{\leq}MMI$$MMI=2.44{\times}log_{10}(PGV(kine))+4.86$ ($I{\leq}MMI$$MMI=2.59{\times}log_{10}(BSPGA(gal{\cdot}sec))-1.02$ ($I{\leq}MMI$

• Journal of Korean Society of Forest Science
• /
• v.101 no.3
• /
• pp.374-380
• /
• 2012

The k-Nearest Neighbor (kNN) technique is popularly applied to assess forest resources at the county level and to provide its spatial information by combining large area forest inventory data and remote sensing data. In this study, two approaches such as distance-weighting and stratification of training dataset, were compared to improve kNN-based forest growing stock estimates. When compared with five distance weights (0 to 2 by 0.5), the accuracy of kNN-based estimates was very similar ranged ${\pm}0.6m^3/ha$ in mean deviation. The training dataset were stratified by horizontal reference area (HRA) and forest cover type, which were applied by separately and combined. Even though the accuracy of estimates by combining forest cover type and HRA- 100 km was slightly improved, that by forest cover type was more efficient with sufficient number of training data. The mean of forest growing stock based kNN with HRA-100 and stratification by forest cover type when k=7 were somewhat underestimated ($5m^3/ha$) compared to statistical yearbook of forestry at 2011.

• Korean Journal of Remote Sensing
• /
• v.37 no.5_1
• /
• pp.871-884
• /
• 2021

The purpose of this study was to determine ways to increase efficiency in constructing and verifying artificial intelligence learning data on land cover using aerial and satellite images, and in applying the data to AI learning algorithms. To this end, multi-resolution datasets of 0.51 m and 10 m each for 8 categories of land cover were constructed using high-resolution aerial images and satellite images obtained from Sentinel-2 satellites. Furthermore, fine data (a total of 17,000 pieces) and coarse data (a total of 33,000 pieces) were simultaneously constructed to achieve the following two goals: precise detection of land cover changes and the establishment of large-scale learning datasets. To secure the accuracy of the learning data, the verification was performed in three steps, which included data refining, annotation, and sampling. The learning data that wasfinally verified was applied to the semantic segmentation algorithms U-Net and DeeplabV3+, and the results were analyzed. Based on the analysis, the average accuracy for land cover based on aerial imagery was 77.8% for U-Net and 76.3% for Deeplab V3+, while for land cover based on satellite imagery it was 91.4% for U-Net and 85.8% for Deeplab V3+. The artificial intelligence learning datasets on land cover constructed using high-resolution aerial and satellite images in this study can be used as reference data to help classify land cover and identify relevant changes. Therefore, it is expected that this study's findings can be used in the future in various fields of artificial intelligence studying land cover in constructing an artificial intelligence learning dataset on land cover of the whole of Korea.

• Journal of Internet Computing and Services
• /
• v.20 no.5
• /
• pp.27-36
• /
• 2019

In this study, experiments on the improvement of the emotion classification, analysis and accuracy of EEG data were proceeded, which applied DEAP (a Database for Emotion Analysis using Physiological signals) dataset. In the experiment, total 32 of EEG channel data measured from 32 of subjects were applied. In pre-processing step, 256Hz sampling tasks of the EEG data were conducted, each wave range of the frequency (Hz); Theta, Slow-alpha, Alpha, Beta and Gamma were then extracted by using Finite Impulse Response Filter. After the extracted data were classified through Time-frequency transform, the data were purified through Independent Component Analysis to delete artifacts. The purified data were converted into CSV file format in order to conduct experiments of Machine learning algorithm and Arousal-Valence plane was used in the criteria of the emotion classification. The emotions were categorized into three-sections; 'Positive', 'Negative' and 'Neutral' meaning the tranquil (neutral) emotional condition. Data of 'Neutral' condition were classified by using Cz(Central zero) channel configured as Reference channel. To enhance the accuracy ratio, the experiment was performed by applying the attributes selected by ASC(Attribute Selected Classifier). In "Arousal" sector, the accuracy of this study's experiments was higher at "32.48%" than Koelstra's results. And the result of ASC showed higher accuracy at "8.13%" compare to the Liu's results in "Valence". In the experiment of Random Forest Classifier adapting ASC to improve accuracy, the higher accuracy rate at "2.68%" was confirmed than Total mean as the criterion compare to the existing researches.

• KIPS Transactions on Software and Data Engineering
• /
• v.10 no.4
• /
• pp.133-142
• /
• 2021

With the recent advancements of deep learning, companies such as smart home, healthcare, and intelligent transportation systems are utilizing its functionality to provide high-quality services for vehicle detection, emergency situation detection, and controlling energy consumption. To provide reliable services in such sensitive systems, deep learning models are required to have high accuracy. In order to develop a deep learning model for analyzing previously mentioned services, developers should utilize the state of the art deep learning models that have already been verified for higher accuracy. The developers can verify the accuracy of the referenced model by validating the model on the dataset. For this validation, the developer needs structural information to document and apply deep learning models, including metadata such as learning dataset, network architecture, and development environments. In this paper, we propose a description language that represents the network architecture of the deep learning model along with its metadata that are necessary to develop a deep learning model. Through the proposed description language, developers can easily verify the accuracy of the referenced deep learning model. Our experiments demonstrate the application scenario of a deep learning description document that focuses on the license plate recognition for the detection of illegally parked vehicles.

• Korean Journal of Radiology
• /
• v.23 no.3
• /
• pp.343-354
• /
• 2022

Objective: To develop and evaluate a deep learning-based artificial intelligence (AI) model for detecting skull fractures on plain radiographs in children. Materials and Methods: This retrospective multi-center study consisted of a development dataset acquired from two hospitals (n = 149 and 264) and an external test set (n = 95) from a third hospital. Datasets included children with head trauma who underwent both skull radiography and cranial computed tomography (CT). The development dataset was split into training, tuning, and internal test sets in a ratio of 7:1:2. The reference standard for skull fracture was cranial CT. Two radiology residents, a pediatric radiologist, and two emergency physicians participated in a two-session observer study on an external test set with and without AI assistance. We obtained the area under the receiver operating characteristic curve (AUROC), sensitivity, and specificity along with their 95% confidence intervals (CIs). Results: The AI model showed an AUROC of 0.922 (95% CI, 0.842-0.969) in the internal test set and 0.870 (95% CI, 0.785-0.930) in the external test set. The model had a sensitivity of 81.1% (95% CI, 64.8%-92.0%) and specificity of 91.3% (95% CI, 79.2%-97.6%) for the internal test set and 78.9% (95% CI, 54.4%-93.9%) and 88.2% (95% CI, 78.7%-94.4%), respectively, for the external test set. With the model's assistance, significant AUROC improvement was observed in radiology residents (pooled results) and emergency physicians (pooled results) with the difference from reading without AI assistance of 0.094 (95% CI, 0.020-0.168; p = 0.012) and 0.069 (95% CI, 0.002-0.136; p = 0.043), respectively, but not in the pediatric radiologist with the difference of 0.008 (95% CI, -0.074-0.090; p = 0.850). Conclusion: A deep learning-based AI model improved the performance of inexperienced radiologists and emergency physicians in diagnosing pediatric skull fractures on plain radiographs.

• The Journal of Advanced Prosthodontics
• /
• v.10 no.1
• /
• pp.8-17
• /
• 2018

PURPOSE. The study aimed to identify the accuracy and reproducibility of preparations made by gypsum materials of various colors using quantitative and semi-quantitative three-dimensional (3D) approach. MATERIALS AND METHODS. A titanium maxillary first molar preparation was created as reference dataset (REF). Silicone impressions were duplicated from REF and randomized into 6 groups (n=8). Gypsum preparations were formed and grouped according to the color of gypsum materials, and light-scanned to obtain prepared datasets (PRE). Then, in terms of accuracy, PRE were superimposed on REF using the best-fit-algorithm and PRE underwent intragroup pairwise best-fit alignment for assessing reproducibility. Root mean square deviation (RMSD) and degrees of similarity (DS) were computed and analyzed with SPSS 20.0 statistical software (${\alpha}=.05$). RESULTS. In terms of accuracy, PREs in 3D directions were increased in the 6 color groups (from 19.38 to $20.88{\mu}m$), of which the marginal and internal variations ranged $51.36-58.26{\mu}m$ and $18.33-20.04{\mu}m$, respectively. On the other hand, RMSD value and DS-scores did not show significant differences among groups. Regarding reproducibility, both RMSD and DS-scores showed statistically significant differences among groups, while RMSD values of the 6 color groups were less than $5{\mu}m$, of which blue color group was the smallest ($3.27{\pm}0.24{\mu}m$) and white color group was the largest ($4.24{\pm}0.36{\mu}m$). These results were consistent with the DS data. CONCLUSION. The 3D volume of the PREs was predisposed towards an increase during digitalization, which was unaffected by gypsum color. Furthermore, the reproducibility of digitalizing scanning differed negligibly among different gypsum colors, especially in comparison to clinically observed discrepancies.

• IE interfaces
• /
• v.24 no.4
• /
• pp.314-322
• /
• 2011

This study analyzes a relationship between Data Envelopment Analysis(DEA) efficiency scores and a normalization index in order to examine the validity of DEA models. A normalization index concerned in this study is 'sales per R&D project fund' which is regarded as a crucial R&D project performance evaluation index in practice. For this correlation analysis, three distinct DEA models are selected such as DEA basic model, DEA/AR-I revised model(i.e. DEA basic model with Acceptance Region Type I constraints) and Super-Efficiency(SE) model. Especially, SE model is adopted where efficient R&D projects(i.e. Decision Making Units, DMU's) with DEA efficiency score of unity from DEA basic model can be further differentiated in ranks. Considering the non-normality and outliers, two rank correlation coefficients such as Spearman's ${\rho}_s$ and Kendall's ${\tau}_B$ are investigated in addition to Pearson's ${\gamma}$. With an up-to-date empirical massive dataset of n = 482 R&D projects associated with R&D Loan Program of Korea Information Communication Promotion Fund in the year of 2011, statistically significant (+) correlations are verified between the normalization index and every model's DEA efficiency scores with all three correlation coefficients. Especially, the congruence verified in this empirical analysis can be a useful reference for enhancing the practitioner's acceptability onto DEA efficiency scores as a real-world R&D project performance evaluation index.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.9 no.7
• /
• pp.2599-2613
• /
• 2015

Multi-index hashing (MIH) is the state-of-the-art method for indexing binary codes, as it di-vides long codes into substrings and builds multiple hash tables. However, MIH is based on the dataset codes uniform distribution assumption, and will lose efficiency in dealing with non-uniformly distributed codes. Besides, there are lots of results sharing the same Hamming distance to a query, which makes the distance measure ambiguous. In this paper, we propose a data-oriented multi-index hashing method (DOMIH). We first compute the covariance ma-trix of bits and learn adaptive projection vector for each binary substring. Instead of using substrings as direct indices into hash tables, we project them with corresponding projection vectors to generate new indices. With adaptive projection, the indices in each hash table are near uniformly distributed. Then with covariance matrix, we propose a ranking method for the binary codes. By assigning different bit-level weights to different bits, the returned bina-ry codes are ranked at a finer-grained binary code level. Experiments conducted on reference large scale datasets show that compared to MIH the time performance of DOMIH can be improved by 36.9%-87.4%, and the search accuracy can be improved by 22.2%. To pinpoint the potential of DOMIH, we further use near-duplicate image retrieval as examples to show the applications and the good performance of our method.

• Journal of Korea Multimedia Society
• /
• v.14 no.10
• /
• pp.1229-1237
• /
• 2011

The volume ray-casting method is one of the direct volume rendering methods that produces high-quality images as well as manipulates semi-transparent object. Although the volume ray-casting method produces high-quality image by sampling in the region of interest, its rendering speed is slow since the color acquisition process is complicated for repetitive memory reference and accumulation of sample values. Recently, the GPU-based acceleration techniques are introduced. However, they require pre-processing or additional memory. In this paper, we propose efficient point-primitive based method to overcome complicated computation of GPU ray-casting. It presents semi-transparent objects, however it does not require preprocessing and additional memory. Our method is fast since it generates point-primitives from volume dataset during sampling process and it projects the primitives onto the image plane. Also, our method can easily cope with OTF change because we can add or delete point-primitive in real-time.