• 한국정보처리학회:학술대회논문집
• /
• 한국정보처리학회 2020년도 춘계학술발표대회
• /
• pp.455-457
• /
• 2020

Recent studies demonstrate that deep learning model is easily biased by trained with unbalanced datasets. For example, the deep network can be trained to make a prediction by background feature instead the real target's feature. For those problem, a measurement called leakage was introduced to digitize this tendency. In this paper, we propose augmentation strategy which are used generally in computer vision problem to remedy this bias problem and we showed a simple augmentation methods have a effect to this task with experiments.

• ETRI Journal
• /
• 제44권2호
• /
• pp.327-338
• /
• 2022

Classification models for human action recognition require robust features and large training sets for good generalization. However, data augmentation methods are employed for imbalanced training sets to achieve higher accuracy. These samples generated using data augmentation only reflect existing samples within the training set, their feature representations are less diverse and hence, contribute to less precise classification. This paper presents new data augmentation and action representation approaches to grow training sets. The proposed approach is based on two fundamental concepts: virtual video generation for augmentation and representation of the action videos through robust features. Virtual videos are generated from the motion history templates of action videos, which are convolved using a convolutional neural network, to generate deep features. Furthermore, by observing an objective function of the genetic algorithm, the spatiotemporal features of different samples are combined, to generate the representations of the virtual videos and then classified through an extreme learning machine classifier on MuHAVi-Uncut, iXMAS, and IAVID-1 datasets.

• 한국컴퓨터정보학회논문지
• /
• 제28권9호
• /
• pp.47-54
• /
• 2023

영아는 비언어적 의사 소통 방식인 울음이라는 수단을 사용한다[1]. 하지만 영아의 울음소리를 파악하는 것에는 어려움이 따른다. 영아의 울음소리를 해석하기 위해 많은 연구가 진행되었다[2,3]. 이에 본 논문에서는 다양한 음성 데이터 증강을 통한 3D 특징 벡터를 이용한 영아의 울음소리 분류를 제안한다. 연구에서는 총 5개의 클래스 복통, 하품, 불편함, 배고픔, 피곤함(belly pain, burping, discomfort, hungry, tired)로 분류된 데이터 세트를 사용한다. 데이터들은 5가지 기법(Pitch, Tempo, Shift, Mixup-noise, CutMix)을 사용하여 증강한다. 증강 기법 중에서 Tempo, Shift, CutMix 기법을 적용하였을 때 성능의 향상을 보여주었다. 최종적으로 우수한 데이터 증강 기법들을 동시 적용한 결과 단일 특징 벡터와 오리지널 데이터를 사용한 모델보다 17.75%의 성능 향상을 도출하였다.

• International Journal of Fuzzy Logic and Intelligent Systems
• /
• 제12권1호
• /
• pp.1-5
• /
• 2012

Data reduction has been used widely in data mining for convenient analysis. Principal component analysis (PCA) and factor analysis (FA) methods are popular techniques. The PCA and FA reduce the number of variables to avoid the curse of dimensionality. The curse of dimensionality is to increase the computing time exponentially in proportion to the number of variables. So, many methods have been published for dimension reduction. Also, data augmentation is another approach to analyze data efficiently. Support vector machine (SVM) algorithm is a representative technique for dimension augmentation. The SVM maps original data to a feature space with high dimension to get the optimal decision plane. Both data reduction and augmentation have been used to solve diverse problems in data analysis. In this paper, we compare the strengths and weaknesses of dimension reduction and augmentation for classification and propose a classification method using data reduction for classification. We will carry out experiments for comparative studies to verify the performance of this research.

• Journal of Mechanical Science and Technology
• /
• 제18권10호
• /
• pp.1819-1828
• /
• 2004

The purpose of this paper is to investigate the effects of slits and swirl vanes on the turbulence augmentation in the flow fields of a gun-type gas burner using an X-type hot-wire probe. The gun-type gas burner adopted in this study is composed of eight slits and swirl vanes located on the surface of an inclined baffle plate. Experiment was carried out at a flow rate of 450 ι/min in burner model installed in the test section of subsonic wind tunnel. Swirl vanes playa role diffusing main flow more remarkably toward the radial direction than axial one, but slits show a reverse feature. Consequently, both slits and swirl vanes remarkably increase turbulence intensity in the whole range of a gun-type gas burner with a cone-type baffle plate.

• 대한두개안면성형외과학회지
• /
• 제17권1호
• /
• pp.20-24
• /
• 2016

Background: Midface concavity is a relatively common facial feature in East Asian populations. Paranasal augmentation is becoming an increasingly popular procedure for patients with mild concavity and normal occlusion. In this study, we evaluate clinical outcomes following a series of paranasal augmentation. Methods: A retrospective review was performed for patients with Class I occlusion who had undergone bilateral paranasal augmentation using custom-made silicone implants, between October 2005 and September 2013. Patient charts were reviewed for demographic information, concomitant operations, and postoperative complications. Preoperative and postoperative (1-month) photographs were used to evaluate operative outcome. Results: The review identified a total of 93 patients meeting study criteria. Overall, aesthetic outcomes were satisfactory. Five-millimeter thick silicone implant was used in 81 cases, and the mean augmentation was 4.26 mm for this thickness. Among the 93 patients, 2 patients required immediate implant removal due to discomfort. An additional 3 patients experienced implant migration without any extrusion. Nine patients complained of transient paresthesia, which had resolved by 2 weeks. There were no cases of hematoma or infection. All patients reported improvement in their lateral profile and were pleased at follow-up. Complications that arose postoperatively included 9 cases of numbness in the upper lip and 3 cases of implant migration. All cases yielded satisfactory results without persisting complications. Sensations were fully restored postoperatively after 1 to 2 weeks. Conclusion: Paranasal augmentation with custom-made silicone implants is a simple, safe, and inexpensive method that can readily improve the lateral profile of a patient with normal occlusion. When combined with other aesthetic procedures, paranasal augmentation can synergistically improve outcome and lead to greater patient satisfaction.

• 한국인터넷방송통신학회논문지
• /
• 제22권5호
• /
• pp.75-86
• /
• 2022

본 연구는 딥러닝 기반 변분 오토인코더(Variational Autoencoder)를 활용하여 수치형 학습데이터 내 클래스 불균형 문제를 해결하고, 학습데이터를 증강하여 학습모델의 성능을 향상시키고자 한다. 우리는 주어진 테이블 데이터에 대하여 인위적으로 레코드 개수를 늘리기 위해 'D-VAE'을 제안한다. 제안 기법은 최적의 데이터 증강을 지원하기 위해 우선 이산화와 특징선택을 수반한 전처리 과정을 수행한다. 이산화 과정에서 k-means 클러스터링을 적용하여 그룹화한 후, 주어진 데이터가 원-핫 인코딩(one-hot encoding) 기법으로 원-핫 벡터(one-hot vector)로 변환한다. 이후, 특징 선택 기법 중 RFECV 기법을 활용하여 예측에 도움이 되는 변수를 가려내고, 이에 대해서만 변분 오토인코더를 활용하여 새로운 학습데이터를 생성한다. 제안 기법의 성능을 검증하기 위해 4가지 유형의 실험 데이터를 활용하여 데이터 증강 비율별로 그 유효성을 입증한다.

• 한국융합학회논문지
• /
• 제8권12호
• /
• pp.55-62
• /
• 2017

개체명 인식(Named Entity Recognition) 시스템은 문서에서 인명(PS), 지명(LC), 단체명(OG)과 같은 개체명을 가지는 단어나 어구를 해당 개체명으로 인식하는 시스템이다. 개체명 인식을 하기위한 전통적인 연구방법으로는 hand-craft된 자질(feature)을 기반으로 모델을 학습하는 통계 기반의 모델이 있다. 최근에는 딥러닝 기반의 RNN(Recurrent Neural Networks), LSTM(Long-short Term Memory)과 같은 모델을 이용하여 문장을 표현하는 자질을 구성하고 이를 개체명 인식과 같이 순서 라벨링(sequence labeling) 문제 해결에 이용한 연구가 제안되었다. 본 연구에서는 한국어 개체명 인식 시스템의 성능 향상을 위해, end-to-end learning 방식이 가능한 딥러닝 기반의 모델에 미리 구축되어 있는 hand-craft된 자질이나 품사 태깅 정보 및 기구축 사전(lexicon) 정보를 추가로 활용하여 자질을 보강(augmentation)하는 방법을 제안한다. 실험 결과 본 논문에서 제안하는 방법에 따라 자질을 보강한 한국어 개체명 인식 시스템의 성능 향상을 확인하였다. 또한 본 연구의 결과를 한국어 자연어처리(NLP) 및 개체명 인식 시스템을 연구하는 연구자들과의 향후 협업 연구를 위해 github를 통해 공개하였다.

• 한국멀티미디어학회논문지
• /
• 제21권10호
• /
• pp.1150-1161
• /
• 2018

Classification is widely used in medical images to categorize patients and non-patients. However, conventional classification requires a complex procedure, including some rigid steps such as pre-processing, segmentation, feature extraction, detection, and classification. In this paper, we propose a novel convolutional neural network (CNN), called LeukemiaNet, to specifically classify two different types of leukemia, including acute lymphoblastic leukemia (ALL) and acute myeloid leukemia (AML), and non-cancerous patients. To extend the limited dataset, a PCA color augmentation process is utilized before images are input into the LeukemiaNet. This augmentation method enhances the accuracy of our proposed CNN architecture from 96.9% to 97.2% for distinguishing ALL, AML, and normal cell images.

• International Journal of Computer Science & Network Security
• /
• 제23권12호
• /
• pp.1-12
• /
• 2023

Ensuring the security of Supervisory Control and Data Acquisition (SCADA) and Industrial Control Systems (ICS) is paramount to safeguarding the reliability and safety of critical infrastructure. This paper addresses the significant threat posed by reconnaissance attacks on SCADA/ICS networks and presents an innovative methodology for enhancing their protection. The proposed approach strategically employs imbalance dataset handling techniques, ensemble methods, and feature engineering to enhance the resilience of SCADA/ICS systems. Experimentation and analysis demonstrate the compelling efficacy of our strategy, as evidenced by excellent model performance characterized by good precision, recall, and a commendably low false negative (FN). The practical utility of our approach is underscored through the evaluation of real-world SCADA/ICS datasets, showcasing superior performance compared to existing methods in a comparative analysis. Moreover, the integration of feature augmentation is revealed to significantly enhance detection capabilities. This research contributes to advancing the security posture of SCADA/ICS environments, addressing a critical imperative in the face of evolving cyber threats.