• 방송공학회논문지
• /
• 제13권6호
• /
• pp.962-965
• /
• 2008

2차원 형상기술자는 형상의 특성을 수치화하여 형상의 분류와 비교를 가능하게 하며, 영상 검색 및 3차원 모델 검색 등에 적용되어왔다. 기존에 개발된 기술자들은 한 형상의 외곽선에 해당하는 폐곡선만을 기술하거나 정밀성이 떨어진다는 한계가 있었다. 이에 따라 본 논문에서는 하나 이상의 폐곡선으로 이루어진 다중형상에 적용하기 위한 다중 곡률-단계 공간 (Multi Curvature-Scale Space) 기술자를 제안한다. 구체적으로, 하나의 폐곡선을 기술하는데 뛰어난 곡률-단계공간 기술자를 각 폐곡선에 적용하고, 이와 함께 전체 형상내의 각 폐곡선들의 배치 형태를 수치화하여 전체 형상을 기술한다. 또한, 기술자를 구성하는 각 값의 가중치를 조절할 수 있게 하여 사용자에 따른 유사도의 모호함을 극복할 수 있게 하였다. 제시하는 다양한 실험 결과는 제안하는 기술자의 유용함을 나타낸다.

• 한국측량학회지
• /
• 제30권5호
• /
• pp.435-444
• /
• 2012

웹 환경에서의 지도서비스 및 위치기반서비스를 제공하기 위해서는 다축척 지도 데이터베이스를 구축하여야 하나, 제작과정이 아직까지 수동편집에 의존하는 경우가 많았다. 본 연구에서는 기본 지도 자료로부터 다축척 지도 데이터베이스를 자동으로 구축하기 위한 지도 일반화 기법을 제안하고 이를 실제 지도 데이터에 적용하여 프로토타입의 다축척 지도 데이터셋을 생성하고자 한다. 지도 일반화 기법으로는 선택 및 삭제, 단순화, 병합 등의 연산자를 조합하여 적용하였으며, 각각 연산자의 알고리듬과 파라미터들은 T$\ddot{o}$pfer's radical law, 지도의 최소도화 기준, 시각적 표현정도 등을 종합적으로 고려하여 실험적으로 결정하였다. 목표 축척수준은 1:1,000, 1:5,000, 1:25,000, 1:100,000, 1:500,000 의 5단계로 설정하였으며, 대상이 되는 기본 지도 자료는 도로명주소 전자지도와 수치지형도를 사용하였다.

• 문화기술의 융합
• /
• 제6권4호
• /
• pp.767-776
• /
• 2020

GcGAN은 기하학적 일관성을 유지하며 영상 간 스타일을 변환하는 딥러닝 모델이다. 그러나 GcGAN은 회전이나 반전(flip) 등의 한정적인 기하 변환으로 영상의 형태를 보존하기 때문에 영상의 세밀한 형태 정보를 제대로 유지하지 못하는 단점을 가지고 있다. 그래서 본 연구에서는 이런 단점을 개선한 새로운 영상 간 변환 기법인 MSGcGAN(Multi-Scale GcGAN)을 제안한다. MSGcGAN은 GcGAN을 확장한 모델로서, 다중 스케일의 영상을 동시에 학습하여 스케일 불변 특징을 추출함으로써, 영상의 의미적 왜곡을 줄이고 세밀한 정보를 유지하는 방향으로 영상 간 스타일 변환을 수행한다. 실험 결과에 의하면 MSGcGAN은 GcGAN보다 정량적 정성적 측면에서 모두 우수하였고, 영상의 전체적인 형태 정보를 잘 유지하면서 스타일을 자연스럽게 변환함을 확인할 수 있었다.

• 한국해양공학회지
• /
• 제12권1호
• /
• pp.135-142
• /
• 1998

In this study, the frequency transfer function of motions are predicted from the result of a full-scale seakeeping trials. Because the real sea has the characteristics of multi-directional waves,we compare the results in the one directional waves with ones in the directional waves. For calculation of the frequency transfer function in the directional waves, Takezawa's inverse estimation method was introduced and the frequency ranges were divided into three parts in order to consider following seas. The full-scale seakeeping trials was executed in the south sea of Korea using the stern trawler. Those results show that analysis method of the multi-directional waves is more reliable than that of one directional waves, and confirm the possibility of applying this method to the full-scale seakeeping trials.

• Advances in nano research
• /
• 제5권1호
• /
• pp.1-12
• /
• 2017

In this paper, the critical buckling temperature of single-walled Boron Nitride nanotube (SWBNNT) is estimated using a new nonlocal first-order shear deformation beam theory. The present model is capable of capturing both small scale effect and transverse shear deformation effects of SWBNNT and is based on assumption that the inplane and transverse displacements consist of bending and shear components, in which the bending components do not contribute toward shear forces and, likewise, the shear components do not contribute toward bending moments. Results indicate the importance of the small scale effects in the thermal buckling analysis of Boron Nitride nanotube.

• The Journal of the Acoustical Society of Korea
• /
• 제27권3E호
• /
• pp.77-83
• /
• 2008

In the process of the sonar image textures produced, the orientation and scale factors are very significant. However, most of the related methods ignore the directional information and scale invariance or just pay attention to one of them. To overcome this problem, we apply Gabor wavelet to extract the features of sonar images, which combine the advantages of both the Gabor filter and traditional wavelet function. The mother wavelet is designed with constrained parameters and the optimal parameters will be selected at each orientation, with the help of bandwidth parameters based on the Fisher criterion. The Gabor wavelet can have the properties of both multi-scale and multi-orientation. Based on our experiment, this method is more appropriate than traditional wavelet or single Gabor filter as it provides the better discrimination of the textures and improves the recognition rate effectively. Meanwhile, comparing with other fusion methods, it can reduce the complexity and improve the calculation efficiency.

• 스마트미디어저널
• /
• 제12권9호
• /
• pp.45-59
• /
• 2023

Weakly supervised object localization (WSOL) is a task of localizing an object in an image using only image-level labels. Previous studies have followed the conventional class activation mapping (CAM) pipeline. However, we reveal the current CAM approach suffers from problems which cause original CAM could not capture the complete defects features. This work utilizes a convolutional neural network (CNN) pretrained on image-level labels to generate class activation maps in a multi-scale manner to highlight discriminative regions. Additionally, a vision transformer (ViT) pretrained was treated to produce multi-head attention maps as an auxiliary detector. By integrating the CNN-based CAMs and attention maps, our approach localizes defective regions without requiring bounding box or pixel-level supervision during training. We evaluate our approach on a dataset of apple images with only image-level labels of defect categories. Experiments demonstrate our proposed method aligns with several Object Detection models performance, hold a promise for improving localization.

• Journal of Advanced Marine Engineering and Technology
• /
• 제29권2호
• /
• pp.217-224
• /
• 2005

A new roof edge detection method based on multi level scales of wavelet function is proposed in this paper roof edge and its direction are obtained in this new methods at one time. Besides. scale characteristics of detecting roof edge is analyzed. And a few new methods on fingerprint image pre-processing are described. A method segmenting foreground/background of fingerprint images is proposed, in which Prior estimation of direction field is not required any more. A segmentation method based on multi-scale roof edges is implemented. and the valid scale range of the method is defined. too. And the method is used to segment ridges and valleys in fingerprint images simultaneously The exact direction fields made up of the direction of each point in ridges can be obtained when detecting ridges exactly based on the roof edge detector, in comparison with the traditional coarse estimation of direction fields. Obviously. it will establish a solid foundation for the sequent fingerprint identification.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• 제15권9호
• /
• pp.3186-3203
• /
• 2021

In this paper, a multi-scale local difference directional number (MLDDN) pattern is proposed for pig identification. Firstly, the color images of individual pig are converted into grey images by the most significant bits (MSB) quantization, which makes the grey values have better discrimination. Then, Gabor amplitude and phase responses on different scales are obtained by convoluting the grey images with Gabor masks. Next, by calculating the main difference of local edge directions instead of traditionally edge information, the directional numbers of Gabor amplitude and phase responses are encoded. Finally, the block histograms of the encoded images are concatenated on each scale, and the maximum pooling is adopted on different scales to avoid the high feature dimension. Experimental results on two pigsties show that MLDDN impressively outperforms the other widely used local descriptors.

• Journal of Information Processing Systems
• /
• 제18권6호
• /
• pp.794-802
• /
• 2022

The recovery of reasonable depth information from different scenes is a popular topic in the field of computer vision. For generating depth maps with better details, we present an efficacious monocular depth prediction framework with coordinate attention and feature fusion. Specifically, the proposed framework contains attention, multi-scale and feature fusion modules. The attention module improves features based on coordinate attention to enhance the predicted effect, whereas the multi-scale module integrates useful low- and high-level contextual features with higher resolution. Moreover, we developed a feature fusion module to combine the heterogeneous features to generate high-quality depth outputs. We also designed a hybrid loss function that measures prediction errors from the perspective of depth and scale-invariant gradients, which contribute to preserving rich details. We conducted the experiments on public RGBD datasets, and the evaluation results show that the proposed scheme can considerably enhance the accuracy of depth prediction, achieving 0.051 for log10 and 0.992 for δ<1.25³ on the NYUv2 dataset.