• 한국재료학회지
• /
• 제31권8호
• /
• pp.445-449
• /
• 2021

Zintl compound Mg₃Sb₂ is a promising candidate for efficient thermoelectric material due to its small band gap energy and characteristic electron-crystal phonon-glass behavior. Furthermore, this compound enables fine tuning of carrier concentration via chemical doping for optimizing thermoelectric performance. In this study, nominal compositions of Mg_3.8Sb_2-xTe_x (0 ≤ x ≤ 0.03) are synthesized through controlled melting and subsequent vacuum hot pressing method. X-ray diffraction (XRD) and scanning electron microscopy (SEM) are carried out to investigate phase development and surface morphology during the process. It should be noted that 16 at. % of excessive Mg must be added to the system to compensate for the loss of Mg during melting process. Herein, thermoelectric properties such as Seebeck coefficient, electrical conductivity, and thermal conductivity are evaluated from low to high temperature regimes. The results show that Te substitution at Sb sites effectively tunes the majority carriers from holes to electrons, resulting in a transition from p to n-type. At 873 K, a peak ZT value of 0.27 is found for the specimen Mg_3.8Sb_1.99Te_0.01, indicating an improved ZT value over the intrinsic value.

• 반도체디스플레이기술학회지
• /
• 제20권3호
• /
• pp.6-10
• /
• 2021

Recognition of characters printed or marked on the PCB components from images captured using cameras is an important task in PCB components inspection systems. Previous optical character recognition (OCR) of PCB components typically consists of two stages: character segmentation and classification of each segmented character. However, character segmentation often fails due to corrupted characters, low image contrast, etc. Thus, OCR without character segmentation is desirable and increasingly used via deep neural networks. Typical implementation based on deep neural nets without character segmentation includes convolutional neural network followed by recurrent neural network (RNN). However, one disadvantage of this approach is slow execution due to RNN layers. LPRNet is a segmentation-free character recognition network with excellent accuracy proved in license plate recognition. LPRNet uses a wide convolution instead of RNN, thus enabling fast inference. In this paper, LPRNet was adapted for recognizing characters printed on PCB components with fast execution and high accuracy. Initial training with synthetic images followed by fine-tuning on real text images yielded accurate recognition. This net can be further optimized on Intel CPU using OpenVINO tool kit. The optimized version of the network can be run in real-time faster than even GPU.

• International Journal of Advanced Culture Technology
• /
• 제9권2호
• /
• pp.86-90
• /
• 2021

Sentiment refers to a person's thoughts, opinions, and feelings toward an object. Sentiment analysis is a process of collecting opinions on a specific target and classifying them according to their emotions, and applies to opinion mining that analyzes product reviews and reviews on the web. Companies and users can grasp the opinions of public opinion and come up with a way to do so. Recently, natural language processing models using the Transformer structure have appeared, and Google's BERT is a representative example. Afterwards, various models came out by remodeling the BERT. Among them, the Facebook AI team unveiled the XLM-R (XLM-RoBERTa), an upgraded XLM model. XLM-R solved the data limitation and the curse of multilinguality by training XLM with 2TB or more refined CC (CommonCrawl), not Wikipedia data. This model showed that the multilingual model has similar performance to the single language model when it is trained by adjusting the size of the model and the data required for training. Therefore, in this paper, we study the improvement of Korean sentiment analysis performed using a pre-trained XLM-R model that solved curse of multilinguality and improved performance.

• Current Optics and Photonics
• /
• 제6권3호
• /
• pp.252-259
• /
• 2022

In this paper, a dual-function dynamically tunable metamaterial absorber is proposed. At frequency points of 1.545 THz and 3.21 THz, two resonance peaks with absorption amplitude of 93.8% (peak I) and 99.4% (peak II) can be achieved. By regulating the conductivity of photosensitive silicon with a pump laser, the resonance frequency of peak I switches to 1.525 THz, and that of peak II switches to 2.79 THz. By adjusting the incident polarization angle by rotating the device, absorption amplitude tuning is obtained. By introducing two degrees of regulation freedom, the absorption amplitude modulation and resonant frequency switching are simultaneously realized. More importantly, dynamic and continuous adjustment of the absorption amplitude is obtained at a fixed resonant frequency, and the modulation depth reaches 100% for both peaks. In addition, the sensing property of the proposed MMA was studied while it was used as a refractive index sensor. Compared with other results reported, our device not only has a dual-function tunable characteristic and the highest modulation depth, but also simultaneously possesses fine sensing performance.

• Nuclear Engineering and Technology
• /
• 제54권2호
• /
• pp.644-652
• /
• 2022

During the pyroprocessing operation, various signals can be collected by process monitoring (PM). These signals are utilized to diagnose process states. In this study, feasibility of using PM for nuclear safeguards of electrorefining operation was examined based on the use of machine learning for detecting off-normal operations. The off-normal operation, in this study, is defined as co-deposition of key elements through reduction on cathode. The monitored process signal selected for PM was cathode potential. The necessary data were produced through electrodeposition experiments in a laboratory molten salt system. Model-based cathodic surface area data were also generated and used to support model development. Computer models for classification were developed using a series of recurrent neural network architectures. The concept of transfer learning was also employed by combining pre-training and fine-tuning to minimize data requirement for training. The resulting models were found to classify the normal and the off-normal operation states with a 95% accuracy. With the availability of more process data, the approach is expected to have higher reliability.

• Journal of Multimedia Information System
• /
• 제9권2호
• /
• 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 Multimedia Information System
• /
• 제9권3호
• /
• pp.177-182
• /
• 2022

A deep neural network (DNN) includes variables whose values keep on changing with the training process until it reaches the final point of convergence. These variables are the co-efficient of a polynomial expression to relate to the feature extraction process. In general, DNNs work in multiple 'dimensions' depending upon the number of channels and batches accounted for training. However, after the execution of feature extraction and before entering the SoftMax or other classifier, there is a conversion of features from multiple N-dimensions to a single vector form, where 'N' represents the number of activation channels. This usually happens in a Fully connected layer (FCL) or a dense layer. This reduced 2D feature is the subject of study for our analysis. For this, we have used the FCL, so the trained weights of this FCL will be used for the weight-class correlation analysis. The popular DNN models selected for our study are ResNet-101, VGG-19, and GoogleNet. These models' weights are directly used for fine-tuning (with all trained weights initially transferred) and scratch trained (with no weights transferred). Then the comparison is done by plotting the graph of feature distribution and the final FCL weights.

• Molecules and Cells
• /
• 제44권5호
• /
• pp.292-300
• /
• 2021

Macrophages residing in various tissue types are unique in terms of their anatomical locations, ontogenies, developmental pathways, gene expression patterns, and immunological functions. Alveolar macrophages (AMs) reside in the alveolar lumen of the lungs and serve as the first line of defense for the respiratory tract. The immunological functions of AMs are implicated in the pathogenesis of various pulmonary diseases such as allergic asthma, chronic obstructive pulmonary disorder (COPD), pulmonary alveolar proteinosis (PAP), viral infection, and bacterial infection. Thus, the molecular mechanisms driving the development and function of AMs have been extensively investigated. In this review article, we discuss the roles of granulocyte-macrophage colony-stimulating factor (GM-CSF) and transforming growth factor (TGF)-β in AM development, and provide an overview of the anti-inflammatory and pro-inflammatory functions of AMs in various contexts. Notably, we examine the relationships between the metabolic status of AMs and their development processes and functions. We hope that this review will provide new information and insight into AM development and function.

• 한국인터넷방송통신학회논문지
• /
• 제22권4호
• /
• pp.23-28
• /
• 2022

최근 스마트폰 또는 타블렛 PC와 같은 스마트기기가 정보의 창구 역할을 하게 되면서 다수의 사용자가 웹포털을 통해 웹 뉴스를 소비하는 것이 더욱 중요해졌다. 하지만 인터넷 상에 생성되는 뉴스의 양을 사용자들이 따라가기 힘들며 중복되고 반복되는 폭발하는 뉴스 기사에 오히려 혼란을 야기 시킬 수도 있다. 본 논문에서는 뉴스 포털에서 사용자의 질의로부터 검색된 뉴스후보들 중 KoBART 기반의 문서요약 기술을 활용한 뉴스 추천 시스템을 제안한다. 실험을 통해서 새롭게 수집된 뉴스 데이터를 기반으로 학습한 KoBART의 성능이 사전훈련보다 더욱 우수한 결과를 보여주었으며 KoBART로부터 생성된 요약문을 환용하여 사용자에게 효과적으로 뉴스를 추천하였다.

• 대한의용생체공학회:의공학회지
• /
• 제43권2호
• /
• pp.102-108
• /
• 2022

Purpose: To propose a deep learning-based clinical decision support technique for laryngeal disease on epiglottis, tongue and vocal cords. Materials and Methods: A total of 873 laryngeal endoscopic images were acquired from the PACS database of Pusan N ational University Yangsan Hospital. and VGG16 model was applied with transfer learning and fine-tuning. Results: The values of precision, recall, accuracy and F1-score for test dataset were 0.94, 0.97, 0.95 and 0.95 for epiglottis images, 0.91, 1.00, 0.95 and 0.95 for tongue images, and 0.90, 0.64, 0.73 and 0.75 for vocal cord images, respectively. Conclusion: Experimental results demonstrated that the proposed model have a potential as a tool for decision-supporting of otolaryngologist during manual inspection of laryngeal endoscopic images.