• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.16 no.9
• /
• pp.2904-2926
• /
• 2022

Currently, diabetes is the most common chronic disease in the world, affecting 23.7% of the population in the Kingdom of Saudi Arabia. Diabetes may be the cause of lower-limb amputations, kidney failure and blindness among adults. Therefore, diagnosing the disease in its early stages is essential in order to save human lives. With the revolution in technology, Artificial Intelligence (AI) could play a central role in the early prediction of diabetes by employing Machine Learning (ML) technology. In this paper, we developed a diagnosis system using machine learning models for the detection of type 2 diabetes among adults, through the adoption of two different diabetes datasets: one for training and the other for the testing, to analyze and enhance the prediction accuracy. This work offers an enhanced classification accuracy as a result of employing several pre-processing methods before applying the ML models. According to the obtained results, the implemented Random Forest (RF) classifier offers the best classification accuracy with a classification score of 98.95%.

• The Transactions of the Korean Institute of Power Electronics
• /
• v.27 no.1
• /
• pp.48-55
• /
• 2022

This paper proposes a battery remaining useful life (RUL) prediction method using a deep learning-based EMD-CNN-LSTM hybrid method. The proposed method pre-processes capacity data by applying empirical mode decomposition (EMD) and predicts the remaining useful life using CNN-LSTM. CNN-LSTM is a hybrid method that combines convolution neural network (CNN), which analyzes spatial features, and long short term memory (LSTM), which is a deep learning technique that processes time series data analysis. The performance of the proposed remaining useful life prediction method is verified using the battery aging experiment data provided by the NASA Ames Prognostics Center of Excellence and shows higher accuracy than does the conventional method.

• Journal of the Institute of Electronics Engineers of Korea SD
• /
• v.47 no.5
• /
• pp.1-9
• /
• 2010

In this paper, the novel architecture of Intra-prediction & DCTQ hardware, which can process for the Full HD image($1980{\times}1088$@30fps) in realtime, is proposed. The cycle optimization method for the overall cycle of prediction, transform, scaling, descaling, and reconstruction is proposed. To reduce the cycle in the $4{\times}4$ prediction, the quantization process is performed during the prediction cycle and pre-selection of 2 modes among the 9 modes is performed to reduce the hardware area. To reduce the hardware of $16{\times}16$ and $8{\times}8$ prediction, the sharing logic between 2 prediction is utilized. The proposed architecture can process the 30frame/sec of full HD image in 108 MHz clock and operate 425 cycle for one macroblock.

• Genomics & Informatics
• /
• v.18 no.1
• /
• pp.5.1-5.5
• /
• 2020

Highly pathogenic avian influenza (HPAI) viruses have caused severe respiratory disease and death in poultry and human beings. Although most of the avian influenza viruses (AIVs) are of low pathogenicity and cause mild infections in birds, some subtypes including hemagglutinin H5 and H7 subtype cause HPAI. Therefore, sensitive and accurate subtyping of AIV is important to prepare and prevent for the spread of HPAI. Next-generation sequencing (NGS) can analyze the full-length sequence information of entire AIV genome at once, so this technology is becoming a more common in detecting AIVs and predicting subtypes. However, an analysis pipeline of NGS-based AIV sequencing data, including AIV subtyping, has not yet been established. Here, in order to support the pre-processing of NGS data and its interpretation, we developed a user-friendly tool, named prediction of avian influenza virus subtype (PAIVS). PAIVS has multiple functions that support the pre-processing of NGS data, reference-guided AIV subtyping, de novo assembly, variant calling and identifying the closest full-length sequences by BLAST, and provide the graphical summary to the end users.

• Journal of the Korea Institute of Military Science and Technology
• /
• v.14 no.3
• /
• pp.492-500
• /
• 2011

The airborne backbone network(ABN) provides communication transport services between airborne nodes, surface nodes and satellite nodes. Such ABN is generally constructed with wide-body and high-capacity planes such as AWACS, which can fly long-term along pre-defined flight paths. In this paper, we propose an efficient method to improve routing performances by reconfiguring routing path before link failure based on the prediction of link state with the information of pre-defined backbone nodes' trajectories. Since the proposed method does not need additional information exchange between airborne nodes in order to acknowledge the link failure, it can be effectively used for airborne backbone network with limited bandwidths.

• Journal of Multimedia Information System
• /
• v.9 no.2
• /
• pp.75-86
• /
• 2022

The existing image quality assessment (IQA) datasets have a small number of samples. Some methods based on transfer learning or data augmentation cannot make good use of image quality-related features. A No Reference (NR)-IQA method based on multi-task training and quality awareness is proposed. First, single or multiple distortion types and levels are imposed on the original image, and different strategies are used to augment different types of distortion datasets. With the idea of weak supervision, we use the Full Reference (FR)-IQA methods to obtain the pseudo-score label of the generated image. Then, we combine the classification information of the distortion type, level, and the information of the image quality score. The ResNet50 network is trained in the pre-train stage on the augmented dataset to obtain more quality-aware pre-training weights. Finally, the fine-tuning stage training is performed on the target IQA dataset using the quality-aware weights to predicate the final prediction score. Various experiments designed on the synthetic distortions and authentic distortions datasets (LIVE, CSIQ, TID2013, LIVEC, KonIQ-10K) prove that the proposed method can utilize the image quality-related features better than the method using only single-task training. The extracted quality-aware features improve the accuracy of the model.

• Journal of Information Processing Systems
• /
• v.19 no.5
• /
• pp.673-687
• /
• 2023

Convolutional neural network-based deep learning technology is the most commonly used in image identification, but it requires large-scale data for training. Therefore, application in specific fields in which data acquisition is limited, such as in the military, may be challenging. In particular, the identification of ground weapon systems is a very important mission, and high identification accuracy is required. Accordingly, various studies have been conducted to achieve high performance using small-scale data. Among them, the ensemble method, which achieves excellent performance through the prediction average of the pre-trained models, is the most representative method; however, it requires considerable time and effort to find the optimal combination of ensemble models. In addition, there is a performance limitation in the prediction results obtained by using an ensemble method. Furthermore, it is difficult to obtain the ensemble effect using models with imbalanced classification accuracies. In this paper, we propose a transfer learning-based feature fusion technique for heterogeneous models that extracts and fuses features of pre-trained heterogeneous models and finally, fine-tunes hyperparameters of the fully connected layer to improve the classification accuracy. The experimental results of this study indicate that it is possible to overcome the limitations of the existing ensemble methods by improving the classification accuracy through feature fusion between heterogeneous models based on transfer learning.

• Transactions of the Korean Society of Mechanical Engineers B
• /
• v.39 no.9
• /
• pp.735-742
• /
• 2015

In this paper, we compare and verify the prediction accuracy and feasibility for wind resources on a wind farm using the Weather Research and Forecasting (WRF) model, which is a numerical weather-prediction model. This model is not only able to simulate local weather phenomena, but also does not require automatic weather station (AWS), satellite, or meteorological mast data. To verify the feasibility of WRF to predict the wind resources required from a wind farm pre-feasibility study, we compare and verify measured wind data and the results predicted by WAsP. To do this, we use the Pyeongdae and Udo sites, which are located on the northeastern part of Jeju island. Together with the measured data, we use the results of annual and monthly mean wind speed, the Weibull distribution, the annual energy production (AEP), and a wind rose. The WRF results are shown to have a higher accuracy than the WAsP results. We therefore confirmed that WRF wind resources can be used in wind farm pre-feasibility studies.

• Journal of Animal Reproduction and Biotechnology
• /
• v.36 no.2
• /
• pp.99-105
• /
• 2021

In horse management, the alarm system with sensors in the foaling period enables the breeder can appropriately prepare the time of the parturition. It is important to prevent losses by unpredictable parturition because there are several high risks such as dystocia and the death of foals and mares during foaling. However, unlike analysis in the alarm system that detects specific motions has been widely performed, analysis of classification following specific behavior patterns or number needs to be more organized. Thus, the objective of this study is to classify signs of the specific behaviors of the mares for the prediction of pre-foaling behaviors. Five Thoroughbred mares (9-20 yrs) were randomly selected for observation of the pre-foaling behaviors. The behaviors were monitored for 90 min that was divided into three different periods as 1) from -90 to -60 min, 2) from -60 to -30 min, 3) from -30 min to the time for the discharge of the amniotic fluid, respectively. The behaviors were divided into two different categories as state and frequent behaviors and each specific behavioral pattern for classification was individually described. In the state behaviors, the number of mares in the standing of the foaling group (3.17 ± 0.18^b) at period 3 was significantly higher than the control group (1.67 ± 0.46^a). In contrast, the number of the mares in the eating of the foaling group (1.17 ± 0.34^b) at period 3 was significantly lower than the control group (3.33 ± 0.46^a). In the frequent behaviors, the weaving of the foaling group was significantly higher than the control group, and looking at the belly of the foaling group was significantly lower than the control group. In period 2, defecation, weaving, and lowering the head of the foaling group were significantly higher than the control group, respectively. In period 3, sitting down and standing up, pawing, weaving, and lowering the head in the foaling group were also significantly higher than the control group. In conclusion, the behavior is significantly different in foaling periods, and the prediction of foaling may be feasible by the detection of the pre-foaling behaviors in the mares.

• Journal of Welding and Joining
• /
• v.34 no.2
• /
• pp.11-15
• /
• 2016

FE analyses for weldment of ship structure are required for various reasons such as stress concentration for bead tow, residual stress and distortion after welding, and hydrogen diffusion for prediction of low temperature crack. These analyses should be done by solid element modeling, but most of ship structures are modeled by shell element. If we are able to make solid element in the shell element FE modeling it is easily to solve the requirement for solid elements in weld analysis of large ship structures. As the nodes of solid element cannot take moments from nodes of shell element, these two kinds of element cannot be used in one model by conventional modeling. The PSCM (Perpendicular shell coupling method) can connect shell to solid. This method uses dummy perpendicular shell element for transferring moment from shell to solid. The target of this study is to develop a FE pre-processing system applicable at welding at ship structure by using PSCM. We also suggested glue-contact technique for controlling element numbers and element qualities and applied it between PSCM and solid element in automatic pre-processing system. The FE weldment modeling through developed pre-processing system will have rational stiffness of adjacent regions. Then FE results can be more reliable when turn-over of ship-block with semi-welded state or ECA (Engineering critical assessment) of weldment in a ship-block are analyzed.