• Proceedings of the Korean Institute of Information and Commucation Sciences Conference
• /
• 2022.05a
• /
• pp.216-217
• /
• 2022

Poisson-Boltzmann equation (PBD), which describes the effects of charges inside cells, plays important roles in various disciplinaries including biology. In this presentation, we introduce a ResNet based method to predict solution of PBE. First, we generate solutions of PBE based on FEM. Next, we train networks whose input shape includes location of charge and shape of cell and while output shape includes the electronic potential.

• Proceedings of the Korean Society of Computer Information Conference
• /
• 2023.07a
• /
• pp.19-20
• /
• 2023

본 논문은 손글씨 인식 분야에서 가장 기본적이고 중요한 주제인 손글씨 데이터 세트에 대한 분류 성능을 분석하고 비교하는 것을 목표로 한다. 이를 위해 ResNet-50 모델을 사용하여 MNIST, EMNIST, KMNIST라는 세 가지 대표적인 손글씨 데이터 세트에 대한 분류 작업을 수행한다. 각 데이터 세트의 특징과 도메인, 그리고 데이터 세트 간의 차이와 특징에 대해 다루며, ResNet-50 모델을 학습하고 평가한 분류 성능을 비교하고 결과에 대해 분석한 결과를 제시한다.

• Proceedings of the Korean Society of Crop Science Conference
• /
• 2023.04a
• /
• pp.49-49
• /
• 2023

• Asian Pacific Journal of Cancer Prevention
• /
• v.14 no.3
• /
• pp.1703-1706
• /
• 2013

Lung cancer remains a deadly disease with unsatisfactory overall survival. Resveratrol (Res) has the potential to inhibit growth of several types of cancer such as prostate and colorectal examples. In the current study, we evaluated in vitro and in vivo anticancer efficiency of Res in a xenograft model with A549 cells. Cell inhibition effects of Res were measured by MTT assay. Apoptotis of A549 cells was assessed with reference to caspase-3 activity and growth curves of tumor volume and bodyweight of the mice were measured every two days. In vitro cytotoxicity evaluation indicated Res to exert dose-dependent cell inhibition effects against A549 cells with activation of caspase-3. In vivo evaluation showed Res to effectively inhibit the growth of lung cancer in a dose-dependent manner in nude mice. Therefore, we believe that Res might be a promising phytomedicine for cancer therapy and further efforts are needed to explore this potential therapeutic strategy.

• Proceedings of the Korean Society of Applied Pharmacology
• /
• 1995.04a
• /
• pp.85-85
• /
• 1995

전년도에 ara-C 유도체 중에서 ara-CDP-DL-PTBA가 각종 암세포주에 대한 뛰어난 항암작용이 보고된바 있으며, 이후 시료의 대량확보, Bulk Formulation 및 전임상등의 세가지 분야에 관해 연구를 수행하였다. Ara-CDP-DL-PTBA를 수용액에 현탁, 초음파분쇄후 NICOMP Analysis에 의하여 micellar solution의 입자도를 알아 본 결과 fresh prepared micelles은 11,1mm size가 88.63%이며 50,5mm가 11.37%로 나타나 평균 19,0mm가 되고, Reconstructed micells은 10.9mm size가 99.87%이며 356.1mm가 0.13%로 나타나 평균 11.0mm가 된다. Ara-C와 Ara-CDP-DL-PTBA의 대사작용을 알아본 결과, ara-C는 투여 1시간째 2,850 $\pm$ 450 pmol, 4시간째 450$\pm$190 pmol, 24시간째 30 pmol 이하로 ara-CTP의 혈중 농도가 급격히 감소하는 반면에 ara-CDP-DL-PTBA는 1시간째 650$\pm$120 pmol, 2시간째 1,800$\pm$500 pmol, 24시간째 300$\pm$90 pmol으로 ara-CTP의 혈중 농도가 서서히 감소하였다.

• ETRI Journal
• /
• v.41 no.2
• /
• pp.242-253
• /
• 2019

We devise a layer-wise hint training method to improve the existing hint-based knowledge distillation (KD) training approach, which is employed for knowledge transfer in a teacher-student framework using a residual network (ResNet). To achieve this objective, the proposed method first iteratively trains the student ResNet and incrementally employs hint-based information extracted from the pretrained teacher ResNet containing several hint and guided layers. Next, typical softening factor-based KD training is performed using the previously estimated hint-based information. We compare the recognition accuracy of the proposed approach with that of KD training without hints, hint-based KD training, and ResNet-based layer-wise pretraining using reliable datasets, including CIFAR-10, CIFAR-100, and MNIST. When using the selected multiple hint-based information items and their layer-wise transfer in the proposed method, the trained student ResNet more accurately reflects the pretrained teacher ResNet's rich information than the baseline training methods, for all the benchmark datasets we consider in this study.

• International Journal of Computer Science & Network Security
• /
• v.22 no.7
• /
• pp.301-307
• /
• 2022

Energy consumption has grown alongside dramatic population increases. Statistics show that buildings in particular utilize a significant amount of energy, worldwide. Because of this, building energy prediction is crucial to best optimize utilities' energy plans and also create a predictive model for consumers. To improve energy prediction performance, this paper proposes a ResNet-LSTM model that combines residual networks (ResNets) and long short-term memory (LSTM) for energy consumption prediction. ResNets are utilized to extract complex and rich features, while LSTM has the ability to learn temporal correlation; the dense layer is used as a regression to forecast energy consumption. To make our model more robust, we employed Huber loss during the optimization process. Huber loss obtains high efficiency by handling minor errors quadratically. It also takes the absolute error for large errors to increase robustness. This makes our model less sensitive to outlier data. Our proposed system was trained on historical data to forecast energy consumption for different time series. To evaluate our proposed model, we compared our model's performance with several popular machine learning and deep learning methods such as linear regression, neural networks, decision tree, and convolutional neural networks, etc. The results show that our proposed model predicted energy consumption most accurately.

• Smart Media Journal
• /
• v.11 no.2
• /
• pp.53-62
• /
• 2022

Due to the accelerated global warming phenomenon after industrialization, the frequency of changes in the existing environment and abnormal climate is increasing. Agriculture is an industry that is very sensitive to climate change, and global warming causes problems such as reducing crop yields and changing growing regions. In addition, environmental changes make the growth period of crops irregular, making it difficult for even experienced farmers to easily estimate the growth stage of crops, thereby causing various problems. Therefore, in this paper, we propose a CNN model for estimating the growth stage of crops. The proposed model was a model that modified the pooling layer of ResNet, and confirmed the accuracy of higher performance than the growth stage estimation of the ResNet and DenseNet models.

• Proceedings of the Korea Information Processing Society Conference
• /
• 2022.11a
• /
• pp.559-561
• /
• 2022

인공지능 시대에 들어서면서 개인 맞춤형 환경을 제공하기 위하여 사람의 감정을 인식하고 교감하는 기술이 많이 발전되고 있다. 사람의 감정을 인식하는 방법으로는 얼굴, 음성, 신체 동작, 생체 신호 등이 있지만 이 중 가장 직관적이면서도 쉽게 접할 수 있는 것은 표정이다. 따라서, 본 논문에서는 정확도 높은 얼굴 감정 식별을 위해서 Convolution Block Attention Module(CBAM)의 각 Gate와 Residual Block, Skip Connection을 이용한 Identity- CBAM Module을 제안한다. CBAM의 각 Gate와 Residual Block을 이용하여 각각의 표정에 대한 핵심 특징 정보들을 강조하여 Context 한 모델로 변화시켜주는 효과를 가지게 하였으며 Skip-Connection을 이용하여 기울기 소실 및 폭발에 강인하게 해주는 모듈을 제안한다. AI-HUB의 한국인 감정 인식을 위한 복합 영상 데이터 세트를 이용하여 총 6개의 클래스로 구분하였으며, F1-Score, Accuracy 기준으로 Identity-CBAM 모듈을 적용하였을 때 Vanilla ResNet50, ResNet101 대비 F1-Score 0.4~2.7%, Accuracy 0.18~2.03%의 성능 향상을 달성하였다. 또한, Guided Backpropagation과 Guided GradCam을 통해 시각화하였을 때 중요 특징점들을 더 세밀하게 표현하는 것을 확인하였다. 결과적으로 이미지 내 표정 분류 Task에서 Vanilla ResNet50, ResNet101을 사용하는 것보다 Identity-CBAM Module을 함께 사용하는 것이 더 적합함을 입증하였다.

• Journal of the Korean Magnetics Society
• /
• v.14 no.4
• /
• pp.143-148
• /
• 2004

Effects of residual flux density M$_{res}$ and number of inspection on the detection voltage and flux density B of the gas pipeline were investigated in MFL inspection, which is widely used for the non-destructive inspection in a gas pipeline. A simulation equipment composed of the magnetizer and iron ring attached on an aluminum disc was constructed instead of a huge gas pipeline facility. With this system. the iron ring could be perfectly demagnetized and signals from the bolt screw stuck on the disc could be clearly detected so that the effects of M$_{res}$S and the inspection number on the detection voltage and B of iron ring were effectively investigated. With increasing the number of inspection, M$_{res}$, B of the iron ring and the detection voltage decreased and then kept at constant values while final M$_{res}$ increased with increasing initial M$_{res}$. If inspection condition were kept unchanged, the detection voltage was proportional to the last M$_{res}$ of the iron ring instead of B. This was probably due to magnetic hysteresis of the iron ring inherited from magnetic domain so that consideration on the magnetic hysteresis was inevitable in the analysis of MFL signal from defects of a gas pipeline. A new inspection scheme using the magnetizer with reversed magnetization in the subsequent inspection was proposed from the result that a high detection voltage could be obtained in the first inspection of gas pipeline with positive M$_{res}$.