• KIPS Transactions on Software and Data Engineering
• /
• v.3 no.8
• /
• pp.329-334
• /
• 2014

In this study, a new method that automatically segments trabecular bone for its morphological analysis using micro-computed tomography imaging was proposed. In the proposed method, the bone region was extracted using a threshold value, and the outer boundary of the bone was detected. The sphere of maximum size with the corresponding voxel as the center was obtained by applying the sphere-fitting method to each voxel of the bone region. If this sphere includes the outer boundary of the bone, the voxels included in the sphere are classified as cortical bone; otherwise, they are classified as trabecular bone. The proposed method was applied to images of the distal femurs of 15 mice, and comparative experiments, with results manually divided by a person, were performed. Four morphological parameters-BV/TV, Tb.Th, Tb.Sp, and Tb.N-for the segmented trabecular bone were measured. The results were compared by regression analysis and the Bland-Altman method; BV/TV, Tb.Th, Tb.Sp, and Tb.N were all in the credible range. In addition, not only can the sphere-fitting method be simply implemented, but trabecular bone can also be divided precisely by using the three-dimensional information.

• Journal of the Institute of Electronics Engineers of Korea SP
• /
• v.42 no.2 s.302
• /
• pp.27-36
• /
• 2005

This thesis proposed an effective encryption method for the DCT-based image/video contents and made it possible to operate in a high speed by implementing it as an optimized hardware. By considering the increase in the amount of the calculation in the image/video compression, reconstruction and encryption, an partial encryption was performed, in which only the important information (DC and DPCM coefficients) were selected as the data to be encrypted. As the result, the encryption cost decreased when all the original image was encrypted. As the encryption algorithm one of the multi-mode AES, DES, or SEED can be used. The proposed encryption method was implemented in software to be experimented with TM-5 for about 1,000 test images. From the result, it was verified that to induce the original image from the encrypted one is not possible. At that situation, the decrease in compression ratio was only $1.6\%$. The hardware encryption system implemented in Verilog-HDL was synthesized to find the gate-level circuit in the SynopsysTM design compiler with the Hynix $0.25{\mu}m$ CMOS Phantom-cell library. Timing simulation was performed by Verilog-XL from CadenceTM, which resulted in the stable operation in the frequency above 100MHz. Accordingly, the proposed encryption method and the implemented hardware are expected to be effectively used as a good solution for the end-to-end security which is considered as one of the important problems.

• KIPS Transactions on Software and Data Engineering
• /
• v.3 no.6
• /
• pp.219-230
• /
• 2014

An analysis of dependency relation is a job that determines the governor and the dependent between words in sentence. The dependency relation of predicate is established by patterns and selectional restriction of subcategorization of the predicate. This paper proposes a method of analysis of Korean dependency relation using homograph predicate disambiguated in morphology analysis phase. The disambiguated homograph predicates has each different pattern. Especially reusing a stage transition training dictionary used during tagging POS and homograph, we propose a method of fixing the dependency relation of {noun+postposition, predicate}, and we analyze the accuracy and an effect of homograph for analysis of dependency relation. We used the Sejong Phrase Structured Corpus for experiment. We transformed the phrase structured corpus to dependency relation structure and tagged homograph. From the experiment, the accuracy of dependency relation by disambiguating homograph is 80.38%, the accuracy is increased by 0.42% compared with one of undisambiguated homograph. The Z-values in statistical hypothesis testing with significance level 1% is ${\mid}Z{\mid}=4.63{\geq}z_{0.01}=2.33$. So we can conclude that the homograph affects on analysis of dependency relation, and the stage transition training dictionary used in tagging POS and homograph affects 7.14% on the accuracy of dependency relation.

• KIPS Transactions on Software and Data Engineering
• /
• v.6 no.3
• /
• pp.161-166
• /
• 2017

In recent times, many people have problems of nutritional imbalance; lack or surplus intake of a specific nutrient despite the variety of available foods. Accordingly, the interest in health and diet issues has increased leading to the emergence of various mobile applications. However, most mobile applications only record the user's diet history and show simple statistics and usually provide only general information for healthy diet. It is necessary for users interested in healthy eating to be provided recommendation services reflecting their food interest and providing customized information. Hence, we propose a menu recommendation method which includes calculating the recommended calorie amount based on the user's physical and activity profile to assign to each food group a substitution unit. In addition, our method also analyzes the user's food preferences using food intake history. Thus it satisfies recommended intake unit for each food group by exchanging the user's preferred foods. Also, the excellence of our proposed algorithm is demonstrated through the calculation of precision, recall, health index and the harmonic average of the 3 aforementioned measures. We compare it to another method which considers user's interest and recommended substitution unit. The proposed method provides menu recommendation reflecting interest and personalized health status by which user can improve and maintain a healthy dietary habit.

• Journal of Broadcast Engineering
• /
• v.18 no.6
• /
• pp.835-849
• /
• 2013

Joint Collaborative Team on Video Coding (JCT-VC), which have founded ISO/IEC MPEG and ITU-T VCEG, has standardized High Efficiency Video Coding (HEVC). Standardization of HEVC has started with purpose of twice or more coding performance compared to H.264/AVC. However, flexible and hierarchical coding block and recursive coding structure are problems to overcome of HEVC standard. Many fast encoding algorithms for reducing computational complexity of HEVC encoder have been proposed. However, it is hard to implement a real-time HEVC encoder only with those fast encoding algorithms. In this paper, for implementation of software-based real-time HEVC encoder, data-level parallelism using SIMD instructions and CPU/GPU multi-threading methods are proposed. And we also proposed appropriate operations and functional modules to apply the proposed methods on HM 10.0 software. Evaluation of the proposed methods implemented on HM 10.0 software showed 20-30fps for $832{\times}480$ sequences and 5-10fps for $1920{\times}1080$ sequences, respectively.

• KIPS Transactions on Software and Data Engineering
• /
• v.9 no.5
• /
• pp.169-176
• /
• 2020

Recently, deep neural networks (DNNs) have brought revolutions to many mobile and embedded devices by providing human-level machine intelligence for various applications. However, high inference accuracy of such DNNs comes at high computational costs, and, hence, there have been significant efforts to reduce computational overheads of DNNs either by compressing off-the-shelf models or by designing a new small footprint DNN architecture tailored to resource constrained devices. One notable recent paradigm in designing small footprint DNN models is sharing parameters in several layers. However, in previous approaches, the parameter-sharing techniques have been applied to large deep networks, such as ResNet, that are known to have high redundancy. In this paper, we propose a parameter-sharing method for already parameter-efficient small networks such as ShuffleNetV2. In our approach, small templates are combined with small layer-specific parameters to generate weights. Our experiment results on ImageNet and CIFAR100 datasets show that our approach can reduce the size of parameters by 15%-35% of ShuffleNetV2 while achieving smaller drops in accuracies compared to previous parameter-sharing and pruning approaches. We further show that the proposed approach is efficient in terms of latency and energy consumption on modern embedded devices.

• KIPS Transactions on Software and Data Engineering
• /
• v.2 no.8
• /
• pp.535-542
• /
• 2013

Since case-based reasoning(CBR) has many advantages, it has been used for supporting decision making in various areas including medical checkup, production planning, customer classification, and so on. However, there are several factors to be set by heuristics when designing effective CBR systems. Among these factors, this study addresses the issue of selecting appropriate neighbors in case retrieval step. As the criterion for selecting appropriate neighbors, conventional studies have used the preset number of neighbors to combine(i.e. k of k-nearest neighbor), or the relative portion of the maximum similarity. However, this study proposes to use the absolute similarity threshold varying from 0 to 1, as the criterion for selecting appropriate neighbors to combine. In this case, too small similarity threshold value may make the model rarely produce the solution. To avoid this, we propose to adopt the coverage, which implies the ratio of the cases in which solutions are produced over the total number of the training cases, and to set it as the constraint when optimizing the similarity threshold. To validate the usefulness of the proposed model, we applied it to a real-world target marketing case of an online shopping mall in Korea. As a result, we found that the proposed model might significantly improve the performance of CBR.

• KIPS Transactions on Software and Data Engineering
• /
• v.5 no.11
• /
• pp.527-534
• /
• 2016

The emergent information society requires the talent for creative thinking based on problem-solving skills and comprehensive thinking rather than simple memorization. Therefore, the Korean curriculum has also changed into the direction of the creative thinking through increasing short-answer questions that can determine the overall thinking of the students. However, their scoring results are a little bit inconsistency because scoring short-answer questions depends on the subjective scoring of human raters. In order to alleviate this point, an automated scoring system using a machine learning has been used as a scoring tool in overseas. Linguistically, Korean and English is totally different in the structure of the sentences. Thus, the automated scoring system used in English cannot be applied to Korean. In this paper, we introduce an automated scoring system for Korean short-answer questions using predictability and unanimity. We also verify the practicality of the automatic scoring system through the correlation coefficient between the results of the automated scoring system and those of human raters. In the experiment of this paper, the proposed system is evaluated for constructed-response items of Korean language, social studies, and science in the National Assessment of Educational Achievement. The analysis was used Pearson correlation coefficients and Kappa coefficient. Results of the experiment had showed a strong positive correlation with all the correlation coefficients at 0.7 or higher. Thus, the scoring results of the proposed scoring system are similar to those of human raters. Therefore, the automated scoring system should be found to be useful as a scoring tool.

• KIPS Transactions on Software and Data Engineering
• /
• v.8 no.9
• /
• pp.355-362
• /
• 2019

In the past, when research and development(R&D) resources were absolutely scarce, the so-called 'choice and concentration' strategy of national R&D projects has been persuasive. Under the current situation where various actors such as GRIs(Government-funded Research Institutes) and universities supported by more abundant R&D resources conduct national R&D projects, this strategy cannot be applied without distinction. In order to see how the strategy has worked, this paper analyzes the concentration of research funds allocated to actors performing national R&D projects. Concentration is measured based on the amount of research funds supported by government from 2002 to 2016 using the Theil index to break down the concentration of individual actors in the overall national R&D project. The results from the Theil index were compared with concentrations using the Gini coefficient, a widely known indicator. As a result, the Theil index could be used to analyze the concentration and sub-components' contribution such as universities and GRIs that make up the entire national R&D system. The results also showed GRIs had the highest concentration, followed by universities, but their concentration has been somewhat reduced compared to 10 years ago. On the other hand, small-sized companies have maintained a certain level, although they are not highly concentrated. In other words, universities and GRIs tend to reduce the gap in the allocation of research funds among institutions, while small-sized companies tend to distribute them evenly.

• KIPS Transactions on Software and Data Engineering
• /
• v.8 no.11
• /
• pp.433-440
• /
• 2019

Recently, machine learning and data mining have been used for many disease prediction and diagnosis. Chronic diseases account for about 80% of the total mortality rate and are increasing gradually. In previous studies, the predictive model for chronic diseases use data such as blood glucose, blood pressure, and insulin levels. In this paper, world's first research, verifies the relationship between dyslipidemia and facial characteristics, and develops the predictive model using machine learning based facial characteristics. Clinical data were obtained from 5390 adult Korean men, and using hypertriglyceridemia and facial characteristics data. Hypertriglyceridemia is a measure of dyslipidemia. The result of this study, find the facial characteristics that highly correlated with hypertriglyceridemia. FD_43_143_aD (p<0.0001, Area Under the receiver operating characteristics Curve(AUC)=0.652) is the best indicator of this study. FD_43_143_aD means distance between mandibular. The model based on this result obtained AUC value of 0.662. These results will provide a basis for predicting various diseases with only facial characteristics in the screening stage of disease epidemiology and public health in the future.