• 정보처리학회논문지:소프트웨어 및 데이터공학
• /
• 제2권3호
• /
• pp.205-208
• /
• 2013

일반적으로 이진패턴 변환은 조명 변화에 강인한 특성을 가지므로, 얼굴인식 및 표정인식 분야에 널리 사용되고 있다. 이에, 본 논문에서는 기존의 LDP(Local Directional Pattern)의 텍스처 성분을 개선한 MLDP(Modified LDP) 변환 영상에 2D-PCA(Two-Dimensional Principal Component Analysis) 알고리즘을 결합한 조명변화에 강인한 얼굴인식 방법에 대하여 제안한다. 기존의 LBP(Local Binary Pattern)나 LDP와 같은 이진패턴 변환들이 히스토그램 특징 추출을 위해 주로 사용되는 것과는 다르게, 본 논문에서 제안하는 방법은 MLDP 영상을 2D-PCA 특징추출을 위해 직접 사용한다는 특성을 갖는다. 제안 방법의 성능평가는 PCA(Principal Component Analysis), 2D-PCA 및 가버변환 영상과 LBP를 결합한 알고리즘을 사용하여, 다양한 조명변화 환경에서 구축된 Yale B 및 CMU-PIE 데이터베이스를 이용하여 수행되었다. 실험 결과, MLDP 영상과 2D-PCA를 사용한 제안 방법이 가장 우수한 인식 성능을 보임을 확인하였다.

• 대한원격탐사학회지
• /
• 제23권4호
• /
• pp.273-285
• /
• 2007

대부분의 스테레오 정합 방법은 두 점간의 대응점을 측정하는데 있어 밝기값을 사용하며 잡음의 영향을 받은 화소가 정합에 사용될 경우 정합 성능이 저하된다. 따라서 잡음은 정합 성능을 결정짓는 중요한 역할을 한다. 본 논문에서는 고해상도 위성영상에서 정합 성능을 향상시키기 위해 잡음에 강건한 밝기 필터와 에지 필터를 이용하여 정합하는 방법을 제안한다. Mean, Median, Midpoint, Gaussian 필터와 같은 밝기 필터와 Gradient, Roberts, Prewitt, Sobel, Laplacian 필터와 같은 에지 필터를 사용하였다. 에지필터와 밝기 필터의 성능 평가를 위해 균일 잡음 또는 가우시안 잡음이 첨가된 합성 영상과 위성 영상에 대해 실험을 수행하였고 필터들은 성능에 따라 순위를 정하였다. 밝기 필터와 에지 필터들 중에서 Median 필터와 Sobel 필터가 가장 우수한 성능을 나타낸 반면에 Midpoint 필터와 Laplacian 필터는 가장 저조한 성능을 나타내었다. Ikonos 스테레오 위성영상을 실험 영상으로 사용하였으며 Median 필터와 Sobel 필터를 이용한 정합 방법이 다른 필터 조합을 이용한 정합 방법보다 향상된 정합 결과를 나타내었다.

• Nuclear Engineering and Technology
• /
• 제47권7호
• /
• pp.791-803
• /
• 2015

The effective cross sections (XSs) in the direct whole core calculation code nTRACER are evaluated by the equivalence theory-based resonance-integral-table method using the WIMS-based library as an alternative to the subgroup method. The background XSs, as well as the Dancoff correction factors, were evaluated by the enhanced neutron-current method. A method, with pointwise microscopic XSs on a union-lethargy grid, was used for the generation of resonance-interference factors (RIFs) for mixed resonant absorbers. This method was modified by the intermediate-resonance approximation by replacing the potential XSs for the non-absorbing moderator nuclides with the background XSs and neglecting the resonance-elastic scattering. The resonance-escape probability was implemented to incorporate the energy self-shielding effect in the spectrum. The XSs were improved using the proposed method as compared to the narrow resonance infinite massbased method. The RIFs were improved by 1% in $^{235}U$, 7% in $^{239}Pu$, and >2% in $^{240}Pu$. To account for thermal feedback, a new feature was incorporated with the interpolation of pre-generated RIFs at the multigroup level and the results compared with the conventional resonance-interference model. This method provided adequate results in terms of XSs and k-eff. The results were verified first by the comparison of RIFs with the exact RIFs, and then comparing the XSs with the McCARD calculations for the homogeneous configurations, with burned fuel containing a mixture of resonant nuclides at different burnups and temperatures. The RIFs and XSs for the mixture showed good agreement, which verified the accuracy of the RIF evaluation using the proposed method. The method was then verified by comparing the XSs for the virtual environment for reactor applicationbenchmark pin-cell problem, as well as the heterogeneous pin cell containing burned fuel with McCARD. The method works well for homogeneous, as well as heterogeneous configurations.

• 융합정보논문지
• /
• 제11권9호
• /
• pp.28-34
• /
• 2021

딥러닝을 이용한 객체 검출의 경우 정확도와 실시간성을 모두 요구한다. 그러나, 한정된 자원 환경에서는 수 많은 양의 데이터를 처리하는 딥러닝 모델을 사용하기 어렵다. 이러한 문제 해결을 위해 본 논문에서는 소형임베디드 장치를 위한 객체 검출을 모델을 제안하였다. 일반적인 검출 모델과 달리 사전 학습된 특징 추출기를 제거한 구조를 사용하여 모델 크기를 최소화하였다. 모델의 구조는 경량화된 컨볼루션 블록을 반복해서 쌓는 구조로 설계하였다. 또한, 검출 오버헤드를 줄이기 위해 영역 제안 횟수를 크게 줄였다. 제안하는 모델은 공개 데이터 셋인 PASCAL VOC를 사용하여 학습 및 평가하였다. 모델의 정량적 평가를 위해 검출 분야에서 사용하는 average precision으로 검출 성능을 측정하였다. 그리고 실제 임베디드 장치와 유사한 라즈베리 파이에서 검출 속도를 측정하였다. 실험을 통해 기존 검출 방법 대비 향상된 정확도와 빠른 추론 속도를 달성하였다.

• 한국음향학회지
• /
• 제40권5호
• /
• pp.488-495
• /
• 2021

본 논문에서는 콘포머 기반 한국어 음성인식 시스템을 제안한다. 콘포머는 트랜스포머 모델에 콘볼루션신경망(Convolution Neural Network, CNN) 기능을 보강한 구조이며 광역 정보를 잘 표현할 수 있는 트랜스포머와 지역 정보를 잘 표현할 수 있는 CNN을 결합한 신경망이다. 음성인식 기본 시스템으로 트랜스포모에 기반한 음성인식시스템을 개발하였으며 언어모델로는 Long Short-Term Memory(LSTM)을 사용하였다. 콘포머 기반 음성인식시스템은 트랜스포머 대신에 콘포머를 사용하였고 언어모델로는 트랜스포머를 이용하였다. 성능 평가를 위해 AI-hub에 있는 Electronics and Telecommunications Research Institute(ETRI) 음성코퍼스를 활용하였으며 트랜스포머 기반 음성인식 시스템은 오인식률이 11.8 %이 되었으며 콘포머 기반 음성인식시스템은 오인식률이 5.7 %가 되었다. AI-hub에 있는 다른 영역의 NHN다이퀘스트 음성 코퍼스를 추가해도 유사한 성능이 유지가 되어 제안된 콘포머 음성인식시스템의 유효성을 입증하였다.

• Journal of Information Science Theory and Practice
• /
• 제9권1호
• /
• pp.24-34
• /
• 2021

The popularity of social networking sites (SNS) has facilitated communication between users. The usage of SNS helps users in their daily life in various ways such as sharing of opinions, keeping in touch with old friends, making new friends, and getting information. However, some users misuse SNS to belittle or hurt others using profanities, which is typical in cyberbullying incidents. Thus, in this study, we aim to identify profane words from the ASKfm corpus to analyze the profane word distribution across four different roles involved in cyberbullying based on lexicon dictionary. These four roles are: harasser, victim, bystander that assists the bully, and bystander that defends the victim. Evaluation in this study focused on occurrences of the profane word for each role from the corpus. The top 10 common words used in the corpus are also identified and represented in a graph. Results from the analysis show that these four roles used profane words in their conversation with different weightage and distribution, even though the profane words used are mostly similar. The harasser is the first ranked that used profane words in the conversation compared to other roles. The results can be further explored and considered as a potential feature in a cyberbullying detection model using a machine learning approach. Results in this work will contribute to formulate the suitable representation. It is also useful in modeling a cyberbullying detection model based on the identification of profane word distribution across different cyberbullying roles in social networks for future works.

• 대한유전성대사질환학회지
• /
• 제18권3호
• /
• pp.95-98
• /
• 2018

미토콘드리아 질환은 단일 장기에서부터 여러 장기에 걸쳐 침범할 수 있다는 임상 증상의 광범위한 이질성이 특징이다. 안검하수, 색소 망막 퇴화, 외안근 마미, 시신경 위축 등과 같은 다양한 안구 증상이 미토콘드리아 질환에서 함께 나타날 수 있지만, 진행성 양안 백내장은 미토콘드리아 질환의 안과적 증상에서 매우 드물다. 저자들은 미토콘드리아 호흡 연쇄 복합체 결핍 환자에서 흔치 않은 안구 발현 현상인 진행성 양안 백내장 침범 사례를 경험하였기에 보고하는 바이다.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• 제13권6호
• /
• pp.3199-3218
• /
• 2019

Location-based services (LBSs) have become popular in recent years due to the ever-increasing usage of smart mobile devices and mobile applications through networks. Although LBS application provides great benefits to mobile users, it also raises a sever privacy concern of users due to the untrusted service providers. In the lack of privacy enhancing mechanisms, most applications of the LBS may discourage the user's acceptance of location services in general, and endanger the user's privacy in particular. Therefore, it is a great interest to discuss on the recent privacy-preserving mechanisms in LBSs. Many existing location-privacy protection-mechanisms (LPPMs) make great efforts to increase the attacker's uncertainty on the user's actual whereabouts by generating a multiple of fake-locations together with user's actual positions. In this survey, we present a study and analysis of existing LPPMs and the state-of-art privacy measures in service quality aware LBS applications. We first study the general architecture of privacy qualification system for LBSs by surveying the existing framework and outlining its main feature components. We then give an overview of the basic privacy requirements to be considered in the design and evaluation of LPPMs. Furthermore, we discuss the classification and countermeasure solutions of existing LPPMs for mitigating the current LBS privacy protection challenges. These classifications include anonymization, obfuscation, and an encryption-based technique, as well as the combination of them is called a hybrid mechanism. Finally, we discuss several open issues and research challenges based on the latest progresses for on-going LBS and location privacy research.

• Journal of Trauma and Injury
• /
• 제32권1호
• /
• pp.40-46
• /
• 2019

Purpose: We perform an analysis of infection risk factors for fracture patients and confirm that the risk factors reported in previous studies increase the risk of actual infection among fractured patients. In addition, injury severity score (ISS) which is used as an evaluation tool for morbidity of trauma patients, confirms whether there is a relationship with infection after orthopedic fracture surgery. Methods: We retrospectively reviewed 1,818 patients who underwent fixation surgery at orthopedic trauma team, focused trauma center from January 1, 2015 to December 31, 2017. Thirty-five patients were infected after fracture surgery. We analyzed age, sex, open fracture criteria based on Gustilo-Aderson classification 3b, anatomical location (upper extremity or lower extremity) of fracture, diabetes, smoking, ISS. Results: Of 1,818 patients, 35 (1.9%) were diagnosed with postoperative infection. Of the 35 infected patients, nine (25.7%) were female and five (14.0%) were upper extremity fractures. Three (8.6%) were diagnosed with diabetes and eight (22.8%) were smokers. Thirteen (37.1%) had ISS less than nine points and six (17.1%) had ISS 15 points or more. Of 1,818 patients, 80 had open fractures. Surgical site infection were diagnosed in 12 (15.0%) of 80. And nine of 12 were checked with Gustilo-Aderson classification 3b or more. Linear logistic regression analysis was performed using statistical analysis program Stata 15 (Stata Corporation, College Station, TX, USA). In addition, independent variables were logistic regression analyzed individually after Propensity scores matching. In all statistical analyzes, only open fracture was identified as a risk factor. Conclusions: The risk factors for infection in fracture patients were found to be significantly influenced by open fracture rather than the underlying disease or anatomical feature of the patient. In the case of ISS, it is considered that there is a limitation. It is necessary to develop a new scoring system that can appropriately approach the morbidity of fracture trauma patients.

• 농업과학연구
• /
• 제45권4호
• /
• pp.823-837
• /
• 2018

The objective of this study was to evaluate the firmness and the sweetness index (SI) of tomatoes with a hyperspectral imaging (HSI) technique within the wavelength range of 1000 - 1550 nm. The hyperspectral images of 95 tomatoes were acquired with a push-broom hyperspectral reflectance imaging system, from which the mean spectra of each tomato were extracted from the regions of interest. The reference firmness and sweetness index of the same sample was measured and calibrated with their corresponding spectral data by partial least squares (PLS) regression with different preprocessing methods. The calibration model developed by PLS regression based on the Savitzky-Golay second-derivative preprocessed spectra resulted in a better performance for both the firmness and the SI of the tomatoes compared to models developed by other preprocessing methods. The correlation coefficients ($R_{pred}$) were 0.82, and 0.74 with a standard error of prediction of 0.86 N, and 0.63, respectively. Then, the feature wavelengths were identified using a model-based variable selection method, i.e., variable importance in projection, from the PLS regression analyses. Finally, chemical images were derived by applying the respective regression coefficients on the spectral image in a pixel-wise manner. The resulting chemical images provided detailed information on the firmness and the SI of the tomatoes. The results show that the proposed HSI technique has potential for rapid and non-destructive evaluation of firmness and the sweetness index of tomatoes.