• 스마트미디어저널
• /
• 제9권4호
• /
• pp.17-25
• /
• 2020

In the image object classification problem, low-resolution images may have a negative impact on the classification result, especially when the classification method, such as a convolutional neural network (CNN) model, is trained on a high-resolution (HR) image dataset. In this paper, we analyze the behavior of applying a classical super-resolution (SR) method such as bicubic interpolation, and a deep CNN model such as SRCNN to enhance low-resolution (LR) weeds images used for classification. Using an HR dataset, we first train a CNN model for weeds image classification with a default input size of 128 × 128. Then, given an LR weeds image, we rescale to default input size by applying the bicubic interpolation or the SRCNN model. We analyze these two approaches on the Chonnam National University (CNU) weeds dataset and find that SRCNN is suitable for the image size is smaller than 80 × 80, while bicubic interpolation is convenient for a larger image.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• 제16권10호
• /
• pp.3390-3405
• /
• 2022

Long Short-Term Memory (LSTM) combined with attention mechanism is extensively used to generate semantic sentences of images in image captioning models. However, features of salient regions and spatial information are not utilized sufficiently in most related works. Meanwhile, the LSTM also suffers from the problem of underutilized information in a single time step. In the paper, two innovative approaches are proposed to solve these problems. First, the Synergy-Gated Attention (SGA) method is proposed, which can process the spatial features and the salient region features of given images simultaneously. SGA establishes a gated mechanism through the global features to guide the interaction of information between these two features. Then, the Recurrent Fusion LSTM (RF-LSTM) mechanism is proposed, which can predict the next hidden vectors in one time step and improve linguistic coherence by fusing future information. Experimental results on the benchmark dataset of MSCOCO show that compared with the state-of-the-art methods, the proposed method can improve the performance of image captioning model, and achieve competitive performance on multiple evaluation indicators.

• 대한임베디드공학회논문지
• /
• 제17권6호
• /
• pp.309-317
• /
• 2022

In this paper, we propose a machine learning-based copy-move forgery detection network with dual branches. Because the rotation or scaling operation is frequently involved in copy-move forger, the conventional convolutional neural network is not effectively applied in detecting copy-move tampering. Therefore, we divide the input into rotation-invariant and scaling-invariant features based on the wavelet coefficients. Each of the features is input to different branches having the same structure, and is fused in the combination module. Each branch comprises feature extraction, correlation, and mask decoder modules. In the proposed network, VGG16 is used for the feature extraction module. To check similarity of features generated by the feature extraction module, the conventional correlation module used. Finally, the mask decoder model is applied to develop a pixel-level localization map. We perform experiments on test dataset and compare the proposed method with state-of-the-art tampering localization methods. The results demonstrate that the proposed scheme outperforms the existing approaches.

• 한국의류산업학회지
• /
• 제25권4호
• /
• pp.429-438
• /
• 2023

In the 21st century fashion industry, the rise of digital environments has transformed it into a dynamic medium, expanding the horizons of media utilization. Consequently, digital fashion film has emerged as a pivotal tool for fashion communication. Functioning as a visual expression medium, fashion film animates fashion concepts into immersive moving images. Proficiency in digital fashion communication has become imperative, considering the attributes of fashion media. Notably, the role of creative directors in ensuring coherent communication across diverse fashion media platforms has gained prominence, underscoring the need for systematic fashion education to nurture specialized talent. This study, therefore, devised a comprehensive curriculum amalgamating fashion communication and practical digital media skills, implemented within fashion major courses. Through this approach, students gained experimental media proficiency and explored innovative approaches to crafting fashion films that eloquently convey fashion narratives. The participants were exposed to the entire spectrum of fashion media production, encompassing digital storytelling, fashion film conceptualization, filming techniques, meticulous editing, and adept utilization of special effects technology. The study's pedagogical strategy, characterized by a focused learning trajectory, garnered significant acclaim. In essence, this study holds significance by formulating a curriculum that nurtures the imaginative and pragmatic aptitudes of fashion majors, immersing them in the dynamic realm of rapidly evolving digital fashion films and their integration with fashion content.

• 정보보호학회논문지
• /
• 제33권6호
• /
• pp.1033-1042
• /
• 2023

본 논문은 네트워크에서 침입 행위를 하는 플로우를 탐지하는 네트워크 침입 탐지 시스템을 제안한다. 대다수 연구에 활용되는 데이터세트는 시계열 정보를 포함하고 있지 않으며, 공격 사례가 적은 공격은 샘플 데이터 수가 부족해 탐지율 향상이 어렵다. 하지만 탐지 방안에 대해 연구 결과가 부족한 상황이다. 본 연구에서는 ANN(Artificial Neural Network) 모델과 스택 앙상블 기법을 활용한 선행 연구를 토대로 하였다. 앞서 언급한 문제점을 해결하기 위해 인접 플로우를 활용하여 시계열 정보를 추가하고 희소 공격의 샘플을 강화하여 학습하여 탐지율을 보강하였다.

• 지능정보연구
• /
• 제18권4호
• /
• pp.43-57
• /
• 2012

제조업에 있어서 판매 후 서비스 건수와 내용 등은 향후 서비스 제공을 위한 자원배분의 효율성 증진과 서비스 품질 향상을 위해서도 매우 중요한 정보이다. 따라서 기업들은 향후 발생하는 판매 후 서비스에 대해 정확히 예측하고 그에 따라 적절히 대처하는 능력을 확보할 필요성이 제조업을 중심으로 증가하고 있다. 그러나 실제로 이들 기업들이 활용하고 있는 서비스 수요예측 방법들은 전통적인 통계적인 예측기법이거나, 시뮬레이션을 기반한 기법들이다. 예를 들면, 전통적인 통계적인 예측기법으로는 회귀분석(regression analysis)의 경우, 다양한 제품모델에 대한 판매 후 서비스 발생 패턴이 선형적인 관계가 매우 적음에도 불구하고 선형으로 가정하여 추정한다는 점과 적정한 회귀식을 가정하여야 되며, 이러한 가정이 실제 경영환경에서는 매우 어렵다는 점 등이 기존의 예측기법들의 한계점으로 지적되고 있다. 본 연구에서는 디지털 TV 모델을 생산 판매 하는 A사의 사례연구를 통하여 최근 인공지능연구에서 각광을 받고 있는 사례기반추론(case-based reasoning; CBR) 기법을 활용한 서비스 수요예측 프레임워크를 제안하고자 한다. 또한, 사례기반추론에서 핵심적인 역할 중 하나인 유사 사례추출 방법에 있어서 가장 일반적인 nearest-neighbor 방법 이외의 유사 사례추출 방법을 제안하고자 한다. 특히, 본 연구에서 제안하는 유사 사례추출 방법은 인공신경망(artificial neural network)을 활용한 자기조직화지도(Self-Organizing Maps : SOM) 군집화 기법을 활용한 유사 사례추출 방식으로 이를 활용한 서비스 수요예측 프레임워크에 구현하고, 실제 기업의 판매 후 서비스 데이터를 활용하여 본 연구에서 제안하는 서비스 수요 예측 프레임워크의 유효성을 실증적으로 검증하고자 한다.

• 대한전자공학회:학술대회논문집
• /
• 대한전자공학회 2002년도 ITC-CSCC -1
• /
• pp.417-420
• /
• 2002

Recent works in the area of multimedia studies focus on a wide range of issues from the impact of multimedia on culture to its impact on economics and anything in between. The interconnectedness of the issues raised by this new practice is complicated by the fact that media are rapidly converging: in a very real way, multimedia is becoming a media prism that reflects the way in which media continually influence each other across disciplines and cultural borders. Thus, the impact of multimedia reflects a complicated crossroads where media, human experience, culture and technology converge. An effective design is generally based on shaping aesthetics for function and utility, with an emphasis on ease of use. However, in designing for cyberspace, it is possible to create narratives that challenge the interactor by encoding in the design an instructional aspect that teaches new approaches and forms. Such a design offers an equally aesthetic experience for the interactor as they explore the meaning of the work. This design approach has been used constructively in many applications. The crucial concern is to determine how little or how much information must be presented for the interactor to achieve a suitable level of cognition. This is always a balancing act: too much difficulty will result in interactor frustration and the abandonment of the activity and too little will result in boredom leading to the same negative result In addition, it can be anticipated that the interactor will bring her or his own level of experiential cognition and/or accretion, to the experience providing reflective cognition and/or restructure the learning curve. If the design of the application is outside their present experience, interactors will begin with established knowledge in order to explore the new work. Thus, it may be argued that the interactor explores, learns and cognates simultaneously based on primary experiential cognition. Learning is one of the most important keys to establishing a comfort level in a new media work. Once interactors have learned a new convention, they apply this cognitive knowledge to other new media experiences they may have. Pierre Levy would describe this process as a "new nomadism" that creates "an invisible space of understanding, knowledge, and intellectual power, within which new qualities of being and new ways of fashioning a society will flourish and mutate" (Levy xxv 1997). Thus, navigation itself of offers the interactors the opportunity to both apply and loam new cognitive skills. This suggests that new media narrative strategies are still in the process of developing unique conventions and, as a result, have not reached a level of coherent grammar. This paper intends to explore the cognitive aspects of new media design and in particular, will explore issues related to the design of new media interfaces. The paper will focus on the creation of narrative strategies that engage interactors through loaming curves thus enhancing interactivity.vity.

• International journal of advanced smart convergence
• /
• 제11권1호
• /
• pp.19-27
• /
• 2022

Across the world, 'housing' comprises a significant portion of wealth and assets. For this reason, fluctuations in real estate prices are highly sensitive issues to individual households. In Korea, housing prices have steadily increased over the years, and thus many Koreans view the real estate market as an effective channel for their investments. However, if one purchases a real estate property for the purpose of investing, then there are several risks involved when prices begin to fluctuate. The purpose of this study is to design a real estate price 'return rate' prediction model to help mitigate the risks involved with real estate investments and promote reasonable real estate purchases. Various approaches are explored to develop a model capable of predicting real estate prices based on an understanding of the immovability of the real estate market. This study employs the LSTM method, which is based on artificial intelligence and deep learning, to predict real estate prices and validate the model. LSTM networks are based on recurrent neural networks (RNN) but add cell states (which act as a type of conveyer belt) to the hidden states. LSTM networks are able to obtain cell states and hidden states in a recursive manner. Data on the actual trading prices of apartments in autonomous districts between January 2006 and December 2019 are collected from the Actual Trading Price Disclosure System of the Ministry of Land, Infrastructure and Transport (MOLIT). Additionally, basic data on apartments and commercial buildings are collected from the Public Data Portal and Seoul Metropolitan Government's data portal. The collected actual trading price data are scaled to monthly average trading amounts, and each data entry is pre-processed according to address to produce 168 data entries. An LSTM model for return rate prediction is prepared based on a time series dataset where the training period is set as April 2015~August 2017 (29 months), the validation period is set as September 2017~September 2018 (13 months), and the test period is set as December 2018~December 2019 (13 months). The results of the return rate prediction study are as follows. First, the model achieved a prediction similarity level of almost 76%. After collecting time series data and preparing the final prediction model, it was confirmed that 76% of models could be achieved. All in all, the results demonstrate the reliability of the LSTM-based model for return rate prediction.

• 한국정보과학회논문지:소프트웨어및응용
• /
• 제30권7_8호
• /
• pp.649-658
• /
• 2003

본 연구에서는 ART2 신경회로망의 성능을 개선하기 위한 계층적 구조를 제안하고, 구성된 클러스터에 대하여 적합도(fitness) 선택을 통한 빠르고 효과적인 패턴 분류 모델(HART2)을 제안한다. 본 논문에서 제안하는 신경회로망은 비지도 학습을 통하여 대략적으로 1차 클러스터를 형성하고, 이 각각의 1차 클러스터로 분류된 패턴에 대해 지도학습을 통한 2군 클러스터를 생성하여 패턴을 분류하는 계층적 신경회로망이다. 이 신경회로망을 이용한 패턴분류 과정은 먼저 입력패턴을 1차 클러스터와 비교하여 유사한 몇 개의 1차 클러스터를 적합도에 따라 선택한다. 이때, 입력패턴과 클러스터들간의 상대 측정 거리비에 기반한 적합도 함수를 도입하여 1차 클러스터에 연결된 클러스터들을 Pruning 함으로써 계층적인 네트워크에서의 속도 향상과 정확성을 추구하였다. 마지막으로 입력패턴과 선택된 1차 클러스터에 연결된 2차 클러스터와의 비교를 통해 최종적으로 패턴을 분류하게 된다. 본 논문의 효율성을 검증하기 위하여 22종의 한글 및 영어 글꼴에 대한 숫자 데이타를 다양한 형태로 변형시켜 확장된 테스트 패턴에 대하여 실험해 본 결과 제안된 신경회로망의 패턴 분류 능력의 우수함을 증명하였다

• 대한원격탐사학회지
• /
• 제38권6_2호
• /
• pp.1737-1741
• /
• 2022

고해상도 드론영상과 딥러닝 기술을 결합한 활성산불 감시는 이제 초기단계로 다방면의 연구개발을 필요로 한다. 이 단보에서는 드론영상 산불탐지에 아직 사용되지 않았던 state-of-the-art (SOTA) 모델인 You Only Look Once Version 7 (YOLOv7) 기반의 활성산불 객체탐지를 수행하였으며, 동일한 데이터셋을 사용한 선행연구에 비해 F1점수가 약 0.05 향상된 성과를 거두었다. 향후 우리나라에서도 광역적인 산불감시에 적용될 수 있도록 추가적인 기술 개발이 계속 필요할 것이다.