• 한국컴퓨터정보학회논문지
• /
• 제13권1호
• /
• pp.117-124
• /
• 2008

본 논문에서는 색상정보와 형태정보를 이용한 내용기반 영상 검색방법을 제안한다. 이미지의 한 가지 특징만을 고려한 내용 기반 이미지 검색은 두 가지 이상의 특징 정보를 이용했을 때와 비교하여 정확도가 떨어져 성능을 저하시킬 수 있다. 따라서 여러 검색 시스템에서는 색상이나 형태, 질감 등과 같은 이미지의 다양한 특징들을 혼합하여 검색에 이용하고 있다. 본 연구는 각 영상의 Hue값에 대한 색상정보와 CSS(Curvature Scale Space)를 이용한 형태정보를 사용한다. 각 영상들의 특징 정보와 데이터베이스에 저장된 영상들의 특징 정보들을 비교하여 유사도 순위에 따라 후보영상들이 검색된다. 실험 결과 색상정보나 형태정보 한가지의 특징만을 사용한 경우 보다 정확도와 재현율면에서 사용자가 원하는 이미지와 보다 유사한 결과를 검출할 수 있었다.

• 한국실내디자인학회:학술대회논문집
• /
• 한국실내디자인학회 2005년도 춘계학술발표대회 논문집
• /
• pp.255-256
• /
• 2005

For the students who specialize in interior design irrespective of their color susceptibility, it is prerequisite to precede a systematic curriculum in the first place on how to put the color theory to practical use. Therefore, this curriculum has its purpose to let the students study on how to put to practical use of their understanding of color as one of the finishing materials for interior design. The 16-weekschool days has been segmented into a theoretical study and practical training, This thesis has summed up the basic instance of the color theory, the first of the three basicpractical courses and presented its results to the students. In due course of this curriculum, the students were recommended to pick up one photo image of their selection out of which they could make out color palette based on their analysis of the photo image, so that they could study on how to apply their color palette to a three dimensional space. Firstly, through this course, they could experience the process for a color image that they had so far sensed subjectively and vaguely to become an objective inevitable result. Secondly, they studied on the process of how a two-dimensionalcolor image could be applied for a three-dimensional space. It is significant of this curriculum in the sense that the students become prepared through this study course with essential knowledge applicable to various interior designs in their future.

• 한국실내디자인학회:학술대회논문집
• /
• 한국실내디자인학회 2007년도 추계학술발표대회 논문집
• /
• pp.193-194
• /
• 2007

The Japanese 'H' motor company is the one of a few engine makers for F1 racing machine in the world. And it has a tradition and ability to make a humanoid walking robot 'Asimo'. But 'H' motor company is known as a motorcycle maker not a car in Korea. It wanted to reinforce the brand image and identity to Korean consumer. And it needed a powerful marketing tool for the brand image distinguished from other global motor companies. It demanded a stand design to show their powerful and unique identity for surpass other rivals in 2007 Seoul Motor Show. This stand for 'H' motor company in 2007 SMS has an area of 1,250m2 and located between huge domestic motor companies, K and SR. The design was planed to show its unique identity and image, overcoming its relatively small size. I designed a round-shape ceiling structures covered whole booth space for enclosure and vortical space sense. That made a strong brand image by light and sharp structures completely distinguished from other stands. And it has a main logo sign for recognizing the stand from a distance.

• 한국항공우주학회지
• /
• 제47권8호
• /
• pp.590-598
• /
• 2019

현대 사회에서 인공위성은 통신, 영상 등의 분야에서 널리 이용되고 있다. 이 중에서도 인공위성을 통해 획득한 영상은 넓은 지역에 대한 다양한 정보를 담고 있어 농업, 자원개발 및 활용, 군사적 목적 등으로 활용되고 있다. 인공위성의 특성상 영상을 획득할 수 있는 시간이 매우 제한적이므로 주어진 시간 내에 최대한 효율적인 영상획득을 수행하는 것이 중요하다. 이를 위해서 본 연구에서는 인공위성이 영상을 획득하는 데 소요되는 시간 및 자세 기동에 대한 모델링을 수행하고 이를 바탕으로 휴리스틱 평가함수를 이용한 임무설계 알고리즘을 제안하고 수치 시뮬레이션을 통하여 해당 알고리즘의 성능을 검증하였다.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• 제13권7호
• /
• pp.3654-3670
• /
• 2019

This paper proposes an Image Texture Median Filter (ITMF) to analyze and detect Android malware on Drebin datasets. We design a model of "ITMF" combined with Image Processing of Median Filter (MF) to reflect the similarity of the malware binary file block. At the same time, using the MAEVS (Malware Activity Embedding in Vector Space) to reflect the potential dynamic activity of malware. In order to ensure the improvement of the classification accuracy, the above-mentioned features(ITMF feature and MAEVS feature)are studied to train Restricted Boltzmann Machine (RBM) and Back Propagation (BP). The experimental results show that the model has an average accuracy rate of 95.43% with few false alarms. to Android malicious code, which is significantly higher than 95.2% of without ITMF, 93.8% of shallow machine learning model SVM, 94.8% of KNN, 94.6% of ANN.

• International journal of advanced smart convergence
• /
• 제8권1호
• /
• pp.153-158
• /
• 2019

Language and letter are represented by a combination of its signifier and signified. All symblos are commonly calling direct attention to people who are living in the physical world. However, in cyberspace, the image takes control once inhabited by the language. Cyberspace with anonymity and deoent physical nature has something with physical laws. In this space the statue transcends the imaginary realm. Ideology gives a greater connotation by giving a series of regulation to images floating in cyberspace. Even if various media have the same image, the meaning of the image changes depending on which ideology is used. The impact of this on the public is great. In this study, I discuss the ideology in cyberspace where is supposedly wide open to who visits and spreads all the thoughts without suppression and the human body. The main ideas would be who owns the ideology and what does it want to control and how the people would react to the ideology. This paper would eventually head the prototype that visualizes above ideas. Though the interactive media it will also show the subject in the real world is accept the images floating in the cyberspace without any doubts.

• Imaging Science in Dentistry
• /
• 제50권4호
• /
• pp.331-337
• /
• 2020

Purpose: As cone-beam computed tomography (CBCT) has become the most widely used 3-dimensional (3D) imaging modality in the dental field, storage space and costs for large-capacity data have become an important issue. Therefore, if 3D data can be stored at a clinically acceptable compression rate, the burden in terms of storage space and cost can be reduced and data can be managed more efficiently. In this study, a deep learning network for super-resolution was tested to restore compressed virtual CBCT images. Materials and Methods: Virtual CBCT image data were created with a publicly available online dataset (CQ500) of multidetector computed tomography images using CBCT reconstruction software (TIGRE). A very deep super-resolution (VDSR) network was trained to restore high-resolution virtual CBCT images from the low-resolution virtual CBCT images. Results: The images reconstructed by VDSR showed better image quality than bicubic interpolation in restored images at various scale ratios. The highest scale ratio with clinically acceptable reconstruction accuracy using VDSR was 2.1. Conclusion: VDSR showed promising restoration accuracy in this study. In the future, it will be necessary to experiment with new deep learning algorithms and large-scale data for clinical application of this technology.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• 제14권11호
• /
• pp.4443-4462
• /
• 2020

This paper proposes a new framework of reversible data hiding scheme for absolute moment truncation coding (AMBTC) compressed images. AMBTC-based RDH can be applied to optical remote sensing (ORS) image transmission, which achieves target region preservation and image compression simultaneously. Existing methods can be concluded as two types. In type I schemes, stego codes mimic the original AMBTC format where no file bloat occurs, yet the carried secret data is limited. Type II schemes utilize predication errors to recode quantity levels of AMBTC codes which achieves significant increase in embedding capacity. However, such recoding causes bloat inside stego format, which is not appropriate in mentioned ORS transmission. The proposed method is a novel type I RDH method which prevents bloat inside AMBTC stego codes with significant improvement in embedding capacity. The AMBTC compressed trios are grouped into two categories according to a given threshold. In smooth trio, the modified low quantity level is constructed by concatenating Huffman codes and secret bits. The reversible contrast mapping (RCM) is performed to complex trios for data embedment. Experiments show that the proposed scheme provides highest payload compared with existing type I methods. Meanwhile, no expansion inside stego codes is caused.

• Journal of Multimedia Information System
• /
• 제8권2호
• /
• pp.101-110
• /
• 2021

In computer vision, single-image super resolution has been an area of research for a significant period. Traditional techniques involve interpolation-based methods such as Nearest-neighbor, Bilinear, and Bicubic for image restoration. Although implementations of convolutional neural networks have provided outstanding results in recent years, efficiency and single model multi-scalability have been its challenges. Furthermore, previous works haven't placed enough emphasis on real-number scalability. Interpolation-based techniques, however, have no limit in terms of scalability as they are able to upscale images to any desired size. In this paper, we propose a convolutional neural network possessing the advantages of the interpolation-based techniques, which is also efficient, deeming it suitable in practical implementations. It consists of convolutional layers applied on the low-resolution space, post-up-sampling along the end hidden layers, and additional layers on high-resolution space. Up-sampling is applied on a multiple channeled feature map via bicubic interpolation using a single model. Experiments on architectural structure, layer reduction, and real-number scale training are executed with results proving efficient amongst multi-scale learning (including scale multi-path-learning) based models.

• 한국BIM학회 논문집
• /
• 제12권1호
• /
• pp.55-61
• /
• 2022

Recently, demand for designing own space is increasing as the rapid growth of home furnishing market. However, there is a limitation that it is not easy to compare the style between before construction view and after view. This study aims to translate real image into another style with GAN model learned with interior images. To implement this, first we established style criteria and collected modern, natural, and classic style images, and experimented with ResNet, UNet, Gradient penalty concept to CycleGAN algorithm. As a result of training, model recognize common indoor image elements, such as floor, wall, and furniture, and suitable color, material was converted according to interior style. On the other hand, the form of furniture, ornaments, and detailed pattern expressions are difficult to be recognized by CycleGAN model, and the accuracy lacked. Although UNet converted images more radically than ResNet, it was more stained. The GAN algorithm allowed us to represent results within 2 seconds. Through this, it is possible to quickly and easily visualize and compare the front and after the interior space style to be constructed. Furthermore, this GAN will be available to use in the design rendering include interior.