• Journal of the Korea Institute of Information and Communication Engineering
• /
• v.23 no.5
• /
• pp.533-539
• /
• 2019

This study proposes a malware classification model that can handle arbitrary length input data using the Microsoft Malware Classification Challenge dataset. We are based on imaging existing data from malware. The proposed model generates a lot of images when malware data is large, and generates a small image of small data. The generated image is learned as time series data by Dynamic RNN. The output value of the RNN is classified into malware by using only the highest weighted output by applying the Attention technique, and learning the RNN output value by Residual CNN again. Experiments on the proposed model showed a Micro-average F1 score of 92% in the validation data set. Experimental results show that the performance of a model capable of learning and classifying arbitrary length data can be verified without special feature extraction and dimension reduction.

• KSCE Journal of Civil and Environmental Engineering Research
• /
• v.39 no.6
• /
• pp.887-892
• /
• 2019

Many researchers have developed a series of vision-based technologies to monitor construction sites automatically. To achieve high performance of vision-based technologies, it is essential to build a large amount and high quality of training image database (DB). To do that, researchers usually visit construction sites, install cameras at the jobsites, and collect images for training DB. However, such human and site-dependent approach requires a huge amount of time and costs, and it would be difficult to represent a range of characteristics of different construction sites and resources. To address these problems, this paper proposes a framework that automatically constructs a training image DB using web crawling techniques. For the validation, the authors conducted two different experiments with the automatically generated DB: construction work type classification and equipment classification. The results showed that the method could successfully build the training image DB for the two classification problems, and the findings of this study can be used to reduce the time and efforts for developing a vision-based technology on construction sites.

• Journal of Korean Home Economics Education Association
• /
• v.20 no.2
• /
• pp.61-73
• /
• 2008

The purpose of this study was to improve interest and attitude of student's Technology-Home Economics Curriculum and to recognize the change of recognition on Traditional image through Natural Dyeing and Jogakbo making on the unit of making living goods with fabrics in Technology-Home Economics curriculum on first year high school. The abridged result of study following this: First, in Natural Dyeing and Jogakbo making, teaching-learning plan applying LT cooperative studying and learning materials have been developed. Second, after applying the developed lesson plan, the result showed that the change in learning interest and attitude about Technology-Home Economics curriculum was positively improved. After executing the class on Natural Dyeing and Jogakbo making, the interest and concern about tradition were very positively upgraded through the result analyzing the change of recognition on traditional image. Through these results, The Teaching-learning plan and learning materials would show high possibility of application as educational contents about traditional culture in the field of education. After executing the class on making of Natural dyeing and Jogakbo, the learning interest degree and attitude was very positively upgraded, and the recognition on traditional image had been changed to the active and positive recognition.

• Communications of Mathematical Education
• /
• v.36 no.3
• /
• pp.335-353
• /
• 2022

The purpose of this study is to identify the characteristics and types of errors in the conceptual image of Korean language learners according to the types of terms in mathematics that are the basis for solving mathematical word problems, and to prepare basic data for effective teaching and learning methods in solving the word problems of Korean language learners. To do this, a case study was conducted targeting four Korean language learners to analyze the specific conceptual images of terms registered in curriculum and terms that were not registered in curriculum but used in textbooks. As a result of this study, first, it is necessary to guide Korean language learners by using sufficient visualization material so that they can form appropriate conceptual definitions for terms in school mathematics. Second, it is necessary to understand the specific relationship between the language used in the home of Korean language learners and the conceptual image of terms in school mathematics. Third, it is necessary to pay attention to the passive term, which has difficulty in understanding the meaning rather than the active term. Fourth, even for Korean language learners who do not have difficulties in daily communication, it is necessary to instruct them on everyday language that are not registered in the curriculum but used in math textbooks. Fifth, terms in school mathematics should be taught in consideration of the types of errors that reflect the linguistic characteristics of Korean language learners shown in the explanation of terms. This recognition is expected to be helpful in teaching word problem solving for Korean language learners with different linguistic backgrounds.

• Journal of Digital Contents Society
• /
• v.19 no.6
• /
• pp.1197-1205
• /
• 2018

In this paper, we perform Single Image Super Resolution(SISR) for acquired images of EOTS or IRST from naval combat system. In order to conduct super resolution, we use Generative Adversarial Networks(GANs), which consists of a generative model to create a super-resolution image from the given low-resolution image and a discriminative model to determine whether the generated super-resolution image is qualified as a high-resolution image by adjusting various learning parameters. The learning parameters consist of a crop size of input image, the depth of sub-pixel layer, and the types of training images. Regarding evaluation method, we apply not only general image quality metrics, but feature descriptor methods. As a result, a larger crop size, a deeper sub-pixel layer, and high-resolution training images yield good performance.

• The Journal of Korean Institute of Next Generation Computing
• /
• v.15 no.3
• /
• pp.16-24
• /
• 2019

Recently, research on emotion recognition based on EEG has attracted great interest from human-robot interaction field. In this paper, we propose a method of labeling using image-based EEG topography instead of evaluating emotions through self-assessment and annotation labeling methods used in MAHNOB HCI. The proposed method evaluates the emotion by machine learning model that learned EEG signal transformed into topographical image. In the experiments using MAHNOB-HCI database, we compared the performance of training EEG topography labeling models of SVM and kNN. The accuracy of the proposed method was 54.2% in SVM and 57.7% in kNN.

• Journal of the Korea Fashion and Costume Design Association
• /
• v.23 no.3
• /
• pp.87-98
• /
• 2021

Due to the influence of COVID-19, many changes have been made in the education methods in universities. In respomse, this study intendsto present an efficient learning method by identifying the impact of professor images on learning immersion and the learning satisfaction of classes taught with blended learning for university students majoring in beauty at H University in Seoul. For final analysis, 232 of the 234 questionnaires administered from May 17, 2021 to June 2, 2021 were analyzed. For statistical analysis, SPSS 21.0 was utilized; frequency analysis was conducted to identify demographic characteristics, factor analysis was used to verify the research model, and regression analysis was conducted to verify the hypothesis. First, images of professors have been shown to affect learning immersion. Second, the professor image were shown to affect learning satisfaction. Third, education immersion has been shown to affect educational satisfaction. In order to overcome the limitations of online lectures in universities that suddenly began with onset of COVID-19, it is believed that students' satisfaction can be increased by applying blended learning as a way to improve the quality of classes.

• Proceedings of the Korean Society of Broadcast Engineers Conference
• /
• fall
• /
• pp.74-77
• /
• 2021

최근 딥러닝을 이용하여 객체 인식 학습을 위한 데이터셋을 구축하는데 있어 시간과 인력을 단축하기 위해 인조 데이터를 생성하는 연구가 진행되고 있다. 하지만 실제 환경과 관계없이 임의의 배경에 배치되어 구축된 데이터셋으로 학습된 네트워크를 실제 환경으로 구성된 데이터셋으로 테스트할 경우 인식률이 저조하다. 이에 본 논문에서는 실제 배경 이미지에 객체 이미지를 합성하고, 다양성을 위해 3차원으로 회전하여 증강하는 인조 데이터셋 생성 시스템을 제안한다. 제안된 방법으로 구축된 인조 데이터셋으로 학습한 네트워크와 실제 데이터셋으로 학습된 네트워크의 인식률을 비교한 결과, 인조 데이터셋의 성능이 실제 데이터셋의 성능보다 2% 낮았지만, 인조 데이터셋을 구축하는 시간이 실제 데이터셋을 구축하는 시간보다 약 11배 빨라 시간적으로 효율적인 데이터셋 구축 시스템임을 증명하였다.

• The Journal of the Convergence on Culture Technology
• /
• v.7 no.3
• /
• pp.591-598
• /
• 2021

It's possible to get the word-vector by the statistical SVD or deep-learning CBOW and LSTM methods and theses ones learn the contexts of forward/backward words or the sequence of following words. It's used to analyze the poems by Ki Hyung-do with similar words recommended by the word-vector showing the core images of the poetry. It seems at first sight that the words don't go well with the images but they express the similar style described by the reference words once you look close the contexts of the specific poems. The word-vector can analogize the words having the same relations with the ones between the representative words for the core images of the poems. Therefore you can analyze the poems in depth and in variety with the similarity and analogy operations by the word-vector estimated with the statistical SVD or deep-learning CBOW and LSTM methods.

• The Transactions of the Korea Information Processing Society
• /
• v.5 no.7
• /
• pp.1933-1942
• /
• 1998

분류 벡터 양자화(classified vector quantization: CVQ)〔2의 부코드북을 설계함에 있어서, 경쟁 학습 네트워크〔5〕-〔7〕 는 소속도의 이분법적 표현으로 상당한 소속도를 가지는 벡터들이 학습 과정에 무시되는 경향을 가진다. 이를 개선하기 위해 제안된 퍼지 경쟁 학습 네트워크〔8〕는 각 클러스터가 연속적인 소속도를 가진다는 개념을 도입하여 이와 같은 문제들을 해결했다. 그러나 퍼지 경쟁 학습 네트워크를 CVQ에 적용할 경우, 각 부코드북의 크기를 시행착오로 결정해야 하는 문제점을 여전히 가지고 있으며, 이러한 문제점들의 개선을 위하여 본 논문에서는 수정 퍼지 경쟁 학습 네트워크(modified fuzzy competitive learning network)를 제안한다. 수정 퍼지 경쟁 학습 네트워크는 퍼지 학습 네트워크가 가지는 이 분법적 소속도를 연속적인 소속도로 확장하여, 학습 과정중에 나타날 수 있는 지역 최소점 도달을 억제하였다.