• 디지털융복합연구
• /
• 제19권11호
• /
• pp.17-22
• /
• 2021

본 논문에서는 기존의 인포그래픽 제작방식을 개선하기 위하여 기계학습 기반의 인포그래픽 자동 추천 시스템을 제안하였다. 이 시스템은 복수의 인포그래픽 이미지를 기계학습하는 부분과 사용자의 기초자료 입력만으로 인포그래픽을 인공지능으로 자동 추천하는 부분으로 구성된다. 추천된 인포그랙픽은 라이브러리 형태로 제공되고, 드래그 & 드롭방식으로 추가적인 자료의 입력이 가능하게 된다. 또한, 입력한 자료의 크기에 따라 인포그래픽 이미지가 동적으로 조절되도록 설계하였다. 기계학습 기반의 인포그래픽 자동 추천 과정을 분석한 결과 레이아웃과 키워드에 대한 일치 성공율은 매우 높고, 타입에 대한 일치 성공률은 다소 낮게 나타났다. 추후 인포그래픽 부분별 이미지 타입에 대한 일치 성공률을 향상시키기 위한 연구가 필요할 것이다.

• 한국콘텐츠학회논문지
• /
• 제22권10호
• /
• pp.213-224
• /
• 2022

본 연구는 상품과 생산, 마케팅 등 다양한 측면에서 장점을 갖고 있어 급격하게 비중이 커져가는 니트 갑피를 활용한 스니커즈의 스타일 유형과 디자인 특성을 분석하여 디자인 전략과 상품 개발의 방향을 설정하는데 목적이 있다. 이를 위해 니트 갑피 스니커즈 시장에서 주목받고 있는 브랜드의 상품과 온라인 마켓에서 인기를 끌고 있는 상품 총 924개를 대상으로 스타일 유형과 디자인 특성을 분석하였다. 분석된 연구의 결과는 첫째, 고유성으로, 전통적인 소재와 제작기법을 따르지 않음으로써 니트 갑피 스니커즈만의 고유한 하이브리드적 스타일과 디자인이 정립되고 있다. 둘째, 복합성으로 컬러와 소재에 있어 다양한 조합을 통해 각각의 특성을 살린 멀티플한 이미지를 창출하고 있다. 셋째, 실험성으로 일반 스니커즈에 비해 한층 더 공학적이고 새로운 재료와 기법을 접목하여 과감한 디자인 변화를 꾀하고 있다. 스니커즈 시장이 날로 확대되어 가는 현 시점에서 산업의 경향을 반영한 본 연구를 통해 신발 분야에 대한 다양한 연구 활동과 산업 분야의 브랜드 기획 방향에 기초적인 정보를 제공할 것으로 기대한다.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• 제16권9호
• /
• pp.2991-3007
• /
• 2022

Two dimensional locality preserving projections (2D-LPP) is an improved algorithm of 2D image to solve the small sample size (SSS) problems which locality preserving projections (LPP) meets. It's able to find the low dimension manifold mapping that not only preserves local information but also detects manifold embedded in original data spaces. However, 2D-LPP is simple and elegant. So, inspired by the comparison experiments between two dimensional linear discriminant analysis (2D-LDA) and linear discriminant analysis (LDA) which indicated that matrix based methods don't always perform better even when training samples are limited, we surmise 2D-LPP may meet the same limitation as 2D-LDA and propose a novel matrix exponential method to enhance the performance of 2D-LPP. 2D-MELPP is equivalent to employing distance diffusion mapping to transform original images into a new space, and margins between labels are broadened, which is beneficial for solving classification problems. Nonetheless, the computational time complexity of 2D-MELPP is extremely high. In this paper, we replace some of matrix multiplications with multiple multiplications to save the memory cost and provide an efficient way for solving 2D-MELPP. We test it on public databases: random 3D data set, ORL, AR face database and Polyu Palmprint database and compare it with other 2D methods like 2D-LDA, 2D-LPP and 1D methods like LPP and exponential locality preserving projections (ELPP), finding it outperforms than others in recognition accuracy. We also compare different dimensions of projection vector and record the cost time on the ORL, AR face database and Polyu Palmprint database. The experiment results above proves that our advanced algorithm has a better performance on 3 independent public databases.

• 한국정보통신학회논문지
• /
• 제26권12호
• /
• pp.1816-1825
• /
• 2022

본 논문에서는 다양한 해상도의 카메라를 사용하는 감시 카메라 네트워크에서 각 사람을 추적하는 새로운 접근 방식을 제안한다. 다수의 비겹침 카메라 상에서 사람 추적 시 기존에 사용되던 사람 특징 정보는 다양한 카메라 시야 조건에 쉽게 영향을 받는다. 이러한 한계를 극복하기 위해 제안하는 시스템은 외모 정보와 함께 얼굴 정보를 활용한다. 일반적으로 감시 카메라로 촬영하는 사람 영상은 해상도가 낮은 경우가 많기 때문에 추적을 용이하게 하기 위해 저해상도 얼굴에서도 유용한 특징을 추출할 수 있어야 한다. 제안하는 추적 방식에서 사람 얼굴 특징을 추출하기 위해 탐지된 얼굴을 정면화한 후 텍스쳐 기반의 특징을 추출한다. 또한 감시 카메라에 포착된 얼굴의 크기가 매우 작은 경우 얼굴을 확대하는 초해상도 기법도 함께 활용한다. 공개된 데이터셋인 Dana36을 이용하여 수행한 실험결과를 통해 제안된 알고리즘의 우수한 성능을 보여준다.

• 인간식물환경학회지
• /
• 제23권3호
• /
• pp.277-291
• /
• 2020

Background and objective: The purpose of this study is to investigate the effects of visitors' positive emotions on satisfaction and loyalty with the perception of restorative environment of a healing garden created in an urban agriculture expo. Methods: The psychological indicators to the images of the healing garden were analyzed by the visitors' demographic variables and the three factors of plant cultivation activity level: plant cultivation experience, plant preference, and plant-related event. Results: Between age groups and occupational groups, significant differences were found statistically. The Perceived Restorativeness Scale(PRS) showed significantly differences between age groups in repose, fascination and legibility. The Positive Affect & Negative Affect Schedule(PANAS) showed statistically significant differences between age groups in positive emotions. In addition, we investigated the correlation between the PANAS and the three factors of plant cultivation experience level, the four factors of the PRS, satisfaction and loyalty. The three factors of plant cultivation experience level, the four factors of the PRS, satisfaction and loyalty showed a positive correlation with positive emotions and were inversely correlated with negative emotions significantly. Multiple regression analysis with dummy variables was conducted to examine the effects of plant cultivation activity level, attention restoration, and the PANAS on healing garden visitors' satisfaction and loyalty. As a result, among the four factors of the PRS, fascination and positive affectivity were significant variables that affect healing garden visitors' satisfaction and loyalty. Conclusion: The results indicated that the higher the attention restoration of visitors due to the fascination of the healing garden and the higher their positive affectivity and the more they have plant-related memories, the higher their impact on healing garden visitors' satisfaction and loyalty. Therefore, fascinating natural environments or greenery landscapes like healing gardens where people can contact plants would reduce negative emotions such as anger and sadness but to increase positive emotions such as pleasure, joy and satisfaction.

• 대한두경부종양학회지
• /
• 제38권2호
• /
• pp.7-13
• /
• 2022

Human papillomavirus (HPV) is a causative agent for a subset of oropharyngeal cancer (OPC). The current standard of care (SOC) for locally advanced OPC is 70 Gy definitive radiotherapy (RT) concurrent with cisplatin, which entails significant proportions of acute and late grade 3 or higher toxicities. Accordingly, discovery of favorable prognosis of HPV-related OPC has led to enthusiasm to attenuate subspecialties therapy in multidisciplinary treatment. Diverse deintensification strategies were investigated in multiple phase 2 trials with an assumption that attenuated treatments result in comparable oncologic outcome and less toxicities compared with SOC. Several trials on chemotherapy deintensification revealed that concomitant administration of cisplatin is not to be omitted or substituted for cetuximab without compromising progression-free survival or local control. A transoral robotic surgery (TORS) is investigated as alternative local treatment, but TORS plus SOC or mild deintensified adjuvant RT showed similar toxicities and inferior oncologic outcomes compared with SOC definitive RT or moderately deintensified RT. However, it has been reported that TORS plus deintensified 30-36 Gy adjuvant RT results in excellent outcome and less late toxicity compared with SOC adjuvant RT. Several phase 2 trials reported apparently equivalent progression-free survival and local control and similar adverse effects with moderately deintensified 60 Gy RT compared with SOC 70 Gy RT. Further dose reduction below 60 Gy has been investigated using biology-directed approaches, which use response to induction chemotherapy or metabolic images to triage HPV-positive OPC for deintensified RT. In summary, these trials provide valuable insights for future directions. Available evidence consistently showed that moderately deintensified RT is effective and safe for HPV-positive OPC in both definitive and adjuvant settings. Concurrent cisplatin remains an essential component without which progression-free survival is significantly compromised for advanced HPV-positive OPC. A simple incorporation of TORS to SOC may be detrimental for oncologic outcome without anticipated toxicity reduction. Given the lack of level 1 evidence, it is prudent to curb an unjustified deviation from the current SOC and limit any deintensified strategies to clinical trials and adhere to the current SOC.

• Membrane and Water Treatment
• /
• 제14권1호
• /
• pp.35-45
• /
• 2023

Biofilms significantly affect the performance of wastewater treatment processes in which biodegradability of numerous microorganisms are actively involved, and various technologies have been applied to secure microbial biofilms. Understanding changes in biofilm characteristics by regulating expression of signaling molecules is important to control and regulate biofilms in membrane bioreactor, i.e., biofouling. This study investigated effects of addition of acyl-homoserine lactones (AHL) as a controllable factor for the microbial signaling system on biofilm formation of Pseudomonas aeruginosa PAO1 and multiple strains in membrane bioreactor. The addition of three AHL, i.e., C4-, C6-, and C8-HSL, at a concentration of 200 ㎍/L, enhanced the formation of the PAO1 biofilm and the degree of increases in the biofilm formation of PAO1 were 70.2%, 76.6%, and 72.9%, respectively. The improvement of biofilm formation of individual strains by C4-HSL was an average of 68%, and the microbial consortia increased by approximately 52.1% in the presence of 200 ㎍/L C4-HSL. CLSM images showed that more bacterial cells were present on the membrane surface after the AHL application. In the COMSTAT results, biomass and thickness were increased up to 2.2 times (PAO1) and 1.6 times (multi-strains) by C4-HSL. This study clearly showed that biofilm formation was increased by the application of AHL to individual strain groups, including PAO1 and microbial consortia, and significant increases were observed when 50 or 100 ㎍/L AHL was administered. This suggests that AHL application can improve the biofilm formation of microorganisms, which could yield an enhancement in efficiency of biofilm control, such as in various biofilm reactors including membrane bioreactor and bioflocculent systems in water/wastewater treatment processes.

• 정보처리학회논문지:소프트웨어 및 데이터공학
• /
• 제11권9호
• /
• pp.371-380
• /
• 2022

합성곱 신경망은 이미지와 같은 격자 형태로 배열된 데이터를 다루는데 널리 사용되고 있는 신경망이다. 일반적인 합성곱 신경망은 합성곱층과 완전연결층으로 구성되며 각 층은 비선형활성함수를 포함하고 있다. 본 논문은 합성곱 신경망의 성능을 향상시키기 위해 결합된 파라메트릭 활성함수를 제안한다. 결합된 파라메트릭 활성함수는 활성함수의 크기와 위치를 변환시키는 파라미터를 적용한 파라메트릭 활성함수들을 여러 번 더하여 만들어진다. 여러 개의 크기, 위치를 변환하는 파라미터에 따라 다양한 비선형간격을 만들 수 있으며, 파라미터는 주어진 입력데이터에 의해 계산된 손실함수를 최소화하는 방향으로 학습할 수 있다. 결합된 파라메트릭 활성함수를 사용한 합성곱 신경망의 성능을 MNIST, Fashion MNIST, CIFAR10 그리고 CIFAR100 분류문제에 대해 실험한 결과, 다른 활성함수들보다 우수한 성능을 가짐을 확인하였다.

• International Journal of Computer Science & Network Security
• /
• 제22권5호
• /
• pp.175-181
• /
• 2022

COVID-19 is an acute respiratory syndrome that affects the host's breathing and respiratory system. The novel disease's first case was reported in 2019 and has created a state of emergency in the whole world and declared a global pandemic within months after the first case. The disease created elements of socioeconomic crisis globally. The emergency has made it imperative for professionals to take the necessary measures to make early diagnoses of the disease. The conventional diagnosis for COVID-19 is through Polymerase Chain Reaction (PCR) testing. However, in a lot of rural societies, these tests are not available or take a lot of time to provide results. Hence, we propose a COVID-19 classification system by means of machine learning and transfer learning models. The proposed approach identifies individuals with COVID-19 and distinguishes them from those who are healthy with the help of Deep Visual Embeddings (DVE). Five state-of-the-art models: VGG-19, ResNet50, Inceptionv3, MobileNetv3, and EfficientNetB7, were used in this study along with five different pooling schemes to perform deep feature extraction. In addition, the features are normalized using standard scaling, and 4-fold cross-validation is used to validate the performance over multiple versions of the validation data. The best results of 88.86% UAR, 88.27% Specificity, 89.44% Sensitivity, 88.62% Accuracy, 89.06% Precision, and 87.52% F1-score were obtained using ResNet-50 with Average Pooling and Logistic regression with class weight as the classifier.

• Journal of Periodontal and Implant Science
• /
• 제53권4호
• /
• pp.295-305
• /
• 2023

Purpose: Various methods have been proposed to achieve the nearly complete decontamination of the surface of implants affected by peri-implantitis. We investigated the in vitro debridement efficiency of multiple decontamination methods (Gracey curettes [GC], glycine air-polishing [G-Air], erythritol air-polishing [E-Air] and titanium brushes [TiB]) using a novel spectrophotometric ink-model in 3 different bone defect settings (30°, 60°, and 90°). Methods: Forty-five dental implants were stained with indelible ink and mounted in resin models, which simulated standardised peri-implantitis defects with different bone defect angulations (30°, 60°, and 90°). After each run of instrumentation, the implants were removed from the resin model, and the ink was dissolved in ethanol (97%). A spectrophotometric analysis was performed to detect colour remnants in order to measure the cumulative uncleaned surface area of the implants. Scanning electron microscopy images were taken to assess micromorphological surface changes. Results: Generally, the 60° bone defects were the easiest to debride, and the 30° defects were the most difficult (ink absorption peak: 0.26±0.04 for 60° defects; 0.32±0.06 for 30° defects; 0.27±0.04 for 90° defects). The most effective debridement method was TiB, independently of the bone defect type (TiB vs. GC: P<0.0001; TiB vs. G-Air: P=0.0017; TiB vs. GE-Air: P=0.0007). GE-Air appeared to be the least efficient method for biofilm debridement. Conclusions: T-brushes seem to be a promising decontamination method compared to the other techniques, whereas G-Air was less aggressive on the implant surface. The use of a spectrophotometric model was shown to be a novel but promising assessment method for in vitro ink studies.