• The Transactions of the Korea Information Processing Society
• /
• v.13 no.2
• /
• pp.41-47
• /
• 2024

When there were disparities in performance between models trained in the time and frequency domains, even after conducting an ensemble, we observed that the performance of the ensemble was compromised due to imbalances in the individual model performances. Therefore, this paper proposes a leakage detection technique to enhance the accuracy of pipeline leakage detection through a step-wise learning approach that extracts features from both the time and frequency domains and integrates them. This method involves a two-step learning process. In the Stage 1, independent model training is conducted in the time and frequency domains to effectively extract crucial features from the provided data in each domain. In Stage 2, the pre-trained models were utilized by removing their respective classifiers. Subsequently, the features from both domains were fused, and a new classifier was added for retraining. The proposed transfer learning-based feature fusion technique in this paper performs model training by integrating features extracted from the time and frequency domains. This integration exploits the complementary nature of features from both domains, allowing the model to leverage diverse information. As a result, it achieved a high accuracy of 99.88%, demonstrating outstanding performance in pipeline leakage detection.

• BMB Reports
• /
• v.49 no.2
• /
• pp.71-72
• /
• 2016

Focal cortical dysplasia type II (FCDII) is a focal malformation of the developing cerebral cortex and the major cause of intractable epilepsy. However, since the molecular genetic etiology of FCD has remained enigmatic, the effective therapeutic target for this condition has remained poorly understood. Our recent study on FCD utilizing various deep sequencing platforms identified somatic mutations in MTOR (existing as low as 1% allelic frequency) only in the affected brain tissues. We observed that these mutations induced hyperactivation of the mTOR kinase. In addition, focal cortical expression of mutant MTOR using in utero electroporation in mice, recapitulated the neuropathological features of FCDII, such as migration defect, cytomegalic neuron and spontaneous seizures. Furthermore, seizures and dysmorphic neurons were rescued by the administration of mTOR inhibitor, rapamycin. This study provides the first evidence that brain somatic activating mutations in MTOR cause FCD, and suggests the potential drug target for intractable epilepsy in FCD patients.

• Geomechanics and Engineering
• /
• v.6 no.5
• /
• pp.455-467
• /
• 2014

As the traffic increased, the original capacity of the tunnel has been unable to meet the needs, so it must be expanded. Based on the features of tunnel, the intermittent trough method must be supposed for tunnel enlarging. Under the situation on the buried deep of the tunnel, it could be used the reasonable arch axis model to descript the past covered rock pressure for mechanism calculating of self-bearing arch. Then establish the three-arch combination effectible model for the analysis which is relied on the tunneling enlarging to Chongqing Yu-Zhou tunnel. It has determined the proper width for the intermittent trough in shallow buried tunnel enlarging.

• Proceedings of the Korea Concrete Institute Conference
• /
• 2002.10a
• /
• pp.619-624
• /
• 2002

A prestressed/precast concrete system was used to build the new Asian Olympic Stadium Project in Pusan, Korea. The stadium(mainly intended for cycle racing) is designed for the 2002 Asian Olympic Games and has a seating capacity of 20,000 spectators plus a few private suites. More than 1300 prestressed/precast components were used and they include single columns, primary beams, cantilever beams, double riser stands, and double tees. Especially, a total of 24 cantilever beams is used on the fourth story for the stands and double tees. These 8m long beams are post-tensioned to prevent cracking, to increase their durability and to serve serviceability by vibration. A cantilever section with cast-in-place topping is 800mm wide and 1500mm deep. Cantilever beams are connected to the column with the corbel by cast-in place concrete. Bonded post-tensioning tendons were assembled at the job site. Dead-end anchorages were installed in the end of cantilever beams and live-end anchorage is the opposite of them. This article presents the geometric layouts, design features and so on.

• IEIE Transactions on Smart Processing and Computing
• /
• v.6 no.3
• /
• pp.193-199
• /
• 2017

Instead of using only spatial features from a single frame for person re-identification, a combination of spatial and temporal factors boosts the performance of the system. A recurrent neural network (RNN) shows its effectiveness in generating highly discriminative sequence-level human representations. In this work, we implement RNN, three Long Short Term Memory (LSTM) network variants, and Gated Recurrent Unit (GRU) on Caffe deep learning framework, and we then conduct experiments to compare performance in terms of size and accuracy for person re-identification. We propose using GRU for the optimized choice as the experimental results show that the GRU achieves the highest accuracy despite having fewer parameters than the others.

• The Bulletin of The Korean Astronomical Society
• /
• v.40 no.2
• /
• pp.48.3-49
• /
• 2015

Scaling relations of early-type galaxies (ETG) provide a deep insight into their formation and evolution. Interestingly enough, most relations extending into the dwarf regimes display non-linear or broken-linear features, unlike the linear relations for normal (i.e., intermediate-mass to giant) ETGs only. Here we investigate the mass-size scaling relation of ETGs using a massive database of galaxies from SDSS DR12. We divide ETGs into two groups by the indication of star formation such as colors, and examine their distinction along the mass-size relation. We find that the mass-size distribution of blue, young normal galaxies is in good agreement with that of dwarf ETGs. Our result suggests that blue, young normal ETGs may serve as links between (passive) normal ETGs and dwarfs. We discuss the possibility of blue, young ETGs being progenitors of dwarf ETGs.

• Proceedings of the KIEE Conference
• /
• 2003.11b
• /
• pp.67-70
• /
• 2003

In this paper the main topologies of one-bit full adders, including the most interesting of those recently proposed, are analyzed and compared for speed, power consumption, and power-delay product. The comparison has been performed on circuits, optimized transistor dimension to minimize power-delay product. The investigation has been carried out with properly defined simulation runs on a Cadence environment using a 0.25-${\mu}m$ process, also including the parasitics derived from layout. Performance has been also compared for different supply voltage values. Thus design guidelines have been derived to select the most suitable topology for the design features required. This paper also proposes a novel figure of merit to realistically compare n-bit adders implemented as a chain of one-bit full adders. The results differ from those previously published both for the more realistic simulations carried out and the more appropriate figure of merit used. They show that, except for short chains of blocks or for cases where minimum power consumption is desired, topologies with only pass transistors or transmission gates are not attractive.

• Journal of Korea Multimedia Society
• /
• v.22 no.12
• /
• pp.1347-1355
• /
• 2019

Deep learning approach based on convolution neural network (CNN) has extensively studied in the field of computer vision. However, periocular feature extraction using CNN was not well studied because it is practically impossible to collect large volume of biometric data. This study uses the ResNet model which was trained with the ImageNet dataset. To overcome the problem of insufficient training data, we focused on the training of multi-layer perception (MLP) having simple structure rather than training the CNN having complex structure. It first extracts features using the pretrained ResNet model and reduces the feature dimension by principle component analysis (PCA), then trains a MLP classifier. Experimental results with the public periocular dataset UBIPr show that the proposed method is effective in person authentication using periocular region. Especially it has the advantage which can be directly applied for other biometric traits.

• Journal of the Korean Society for Precision Engineering
• /
• v.24 no.11
• /
• pp.126-133
• /
• 2007

Tuberculosis (TB) is a chronic or acute infectious disease which damages more people than any other infectious diseases according to WHO estimates. In this paper, a new automatic detection system of tuberculosis bacilli by using their color information is proposed. Through the deep investigation of color and intensity compositions of tuberculosis images, new pre-processing and segmentation algorithms are suggested. Specific features of bacilli are extracted from the processed images and number counting is done by using domain-specific knowledge rules.

• Fisheries and Aquatic Sciences
• /
• v.10 no.4
• /
• pp.196-199
• /
• 2007

The spermatozoa of Leiocassis nitidus are relatively simple cells composed of a spherical head, a short midpiece, and a tail, as in most Siluriformes. The ultrastructure is characterized by the following features: Acrosome absent, as in most teleosts; around nucleus about $1.8\;{\mu}m$ long, with a deep nuclear fossa containing the proximal and distal centrioles and mitochondria. Two centrioles approximately $180^{\circ}$ from each other; 10 or more mitochondria surrounding the axoneme (with a 9+2 microtubular pattern), arranged in two layers in the postnuclear cytoplasm and separated from the axoneme by the cytoplasmic canal. Two lateral fins on the same plane as the two central microtubules; doublets 3 and 8, which are ultrastructural characteristics of the sperma tail unlike other siluroids laking the lateral fins.