• Journal of the Optical Society of Korea
• /
• 제18권5호
• /
• pp.507-516
• /
• 2014

We present a hybrid method for spectral reflectivity recovery, using 3D extrapolation as a supplemental method for 3D interpolation. The proposed 3D extrapolation is an extended version of 3D interpolation based on the barycentric algorithm. It is faster and more accurate than the conventional spectral-recovery techniques of principal-component analysis and nonnegative matrix transformation. Four different extrapolation techniques (based on nearest neighbors, circumcenters, in-centers, and centroids) are formulated and applied to recover spectral reflectivity. Under the standard conditions of a D65 illuminant and 1964 $10^{\circ}$ observer, all reflectivity data from 1269 Munsell color chips are successfully reconstructed. The superiority of the proposed method is demonstrated using statistical data to compare coefficients of correlation and determination. The proposed hybrid method can be applied for fast and accurate spectral reflectivity recovery in image processing.

• 제어로봇시스템학회:학술대회논문집
• /
• 제어로봇시스템학회 2001년도 ICCAS
• /
• pp.129.1-129
• /
• 2001

In this paper, we present an approach for matching images from finding interesting points and applying new image interpolation algorithm. New algorithms are developed that automatically align the input images match them and reconstruct 3-D surfaces. The interpolation algorithm is designed to cope with simple shapes. The proposed image interpolation algorithm generate a rotation image about vertical axes by an any angle from 4 base images. Each base image that was obtained from CCD camera has an angle difference of 90$^{\circ}$ The proposed image interpolation algorithm use the geometric analysis of image and depth information.

• 대한기계학회:학술대회논문집
• /
• 대한기계학회 2008년도 추계학술대회A
• /
• pp.1204-1209
• /
• 2008

Design is important in the IT, digital appliance, and auto industries. Aesthetic and art images are being applied for better design satisfaction of the products. Various artistic image patterns are used to satisfy demand of design, but it takes much lead-time and effort to implement them for making dies and molds. In this paper, a hybrid reverse engineering method generating accurate 3D engraving models from 2D art images is proposed through image processing, 3D reconstruction, and NURBS interpolation methods. In order to generate the 3D model from the 2D artistic image, cloud points with z-depth are extracted according to intensity values of the image. An adaptive median filter and harmonic filter are used to obtain the intensity values accurately. NURBS surfaces are generated through the interpolation of the cloud points. Performance of the developed system is to be confirmed through the realization of Mona Lisa and Golden Gate Bridge.

• 대한전자공학회:학술대회논문집
• /
• 대한전자공학회 2000년도 하계종합학술대회 논문집(5)
• /
• pp.138-141
• /
• 2000

In this paper, we present a new interpolation algorithm for three-dimensional images. Generally, Image interpolation is carried out along the three orthogonal coordinates. However, such a interpolation algorithm along orthogonal coordinates do not utilize the contour of 3 dimensional objects. In this paper, we propose a new directional interpolation algorithm that searches the best interpolation direction for 3-dimensional objects. Experiments with brain MR images show promising results.

• 제어로봇시스템학회논문지
• /
• 제8권10호
• /
• pp.874-880
• /
• 2002

A number of techniques have been proposed for 3D display using view-difference of two eyes. These methods do not express enough reality like real world. The display images have to change according to the position of a viewer to improve reality. In this paper, we present an approach for generating intermediate image between two different view images by applying new image interpolation algorithm The interpolation algorithm is designed to cope with complex shapes. The proposed image interpolation algorithm generates rotated image about vertical axes by any angle from base images. Each base image that was obtained from CCD camera has an view-angle difference of $3^{\circ}C$, $5.5^{\circ}C$, $^{\circ}C$, $22^{\circ}C$, and $45^{\circ}C$. The proposed into mediate image generation method uses the geometric analysis of image and depth information through the block-based matching method.

• 한국산업정보학회논문지
• /
• 제9권1호
• /
• pp.7-16
• /
• 2004

효율적인 최종 NC코드를 만들어내기 위해서 공구경로는 효과적인 방법으로 결정되어 져야한다. 이는 3축 CNC 공작기계상에서 자유형상의 벽을 가진 자유곡면을 가공할 때 특히 중요하다. 많은 CAD/CAM 시스템들은 자유곡면가공을 위한 NC코드 생성 시 직선 보간을 이용한다. 그러나 이 방법은 가공된 바닥 표면의 매끄러움과 가공시간, 그리고 CL파일 사이즈를 줄이기 위해서 보정되어야 한다. 곡선 가공이 이러한 문제들을 줄이기 위한 해결책이 될 수 있다. 단일화된 황삭과 정삭 알고리즘과 공구경로가 그래픽으로 시뮬레이션 되었다. 본 연구에서는 자유곡면 바닥을 가진 파켓을 3축 CNC 공작기계를 이용하여 일반적인 파켓 밀링을 위한 직선 보간과 직선+아크를 혼합한 보간을 위한 NC 공구경로 데이터를 만들기 위한 3D 격자 항행 알고리즘을 개발하였다.

• 한국정보통신학회논문지
• /
• 제16권2호
• /
• pp.253-258
• /
• 2012

자연 샘플링 펄스폭 변조 방식을 디지털 영역에서 적용하기 위해서는 왜곡성능이 우수하고 계산 량이 적은 보간 필터가 필수적이다. 이를 위하여 본 논문에서는 변형된 패로우 구조를 기반으로 한 새로운 형태의 보간 필터를 제안 및 설계한다. 제안된 필터는 아날로그 신호 추정을 위해 단지 4개의 주변 샘플 값만을 필요로 하는 단순화된 구조를 가지므로 기존의 라그랑지 보간 필터와 비슷한 복잡도를 유지하면서 더 우수한 통과 대역 왜곡 특성을 보인다. 설계된 보간 필터를 이용하여 20 kHz 의 가청 주파수 대역에서 10-3 dB 의 최대 왜곡 편차를 달성하여 기존의 라그랑지 보간 방식과 비교하여 1/6 정도 감소된 최대 왜곡 편차 특성을 실현하였다.

• 한국멀티미디어학회논문지
• /
• 제14권8호
• /
• pp.1000-1009
• /
• 2011

영상 보간법은 영상의 크기변환에서 할당되지 못한 화소에 대한 값을 추정하는 기술로써, 보간된 결과 영상에서 나타나는 화질 열화 현상을 최소화하면서도 낮은 계산복잡도를 가지는 것이 필요하다. 본 논문에서는 반복적 오차 제거를 이용한 영상 보간법을 제안한다. 제안하는 방법은 5단계로 구성되며, 각각 손실 정보 계산 단계, 손실 정보 추정 단계 손실 정보 적용 단계 오차 계산 단계 오차 적용 단계이다. 실험을 통해서 기존의 방법보다 평균 3.3dB이상 PSNR(peak signal to noise rate)이 향상된 것을 알 수 있었고, 주관적인 화질도 개선된 것을 확인하였으며 계산복잡도가 최소 83% 이상 감소한 것을 측정하였다. 제안한 영상 보간 방법은 영상 복원 및 확대를 위한 다양한 응용 환경에서 유용하게 사용될 수 있다.

• 한국해양공학회:학술대회논문집
• /
• 한국해양공학회 2006년 창립20주년기념 정기학술대회 및 국제워크샵
• /
• pp.327-330
• /
• 2006

With the increasing requirements for the survey and development of the ocean, the demands on the of AUV(Autonomous Underwater Vehicle) technologies have been increased. Reconstruction and replay of the AUV motion on the basis of the data stored during the execution of mission, can help the development of control strategies for AUVs such as mission planning and control algorithms. While an AUV cruises for her mission, her attitude and position data are is recorded. The data can be used for visualization of the motion in off-line. However, because most of the position data gathered from acoustic sensors have long time-interval and include intermittent faulty signal, the replayed motion by the graphic simulator can not demonstrate the motion as a smooth movie. In this paper, interpolation methods are surveyed to reconstruct the AUV position data. Then, an efficient 3D visualization method for AUV motion using the interpolation method is proposed. Simulation results arc also included to verify the proposed method.

• International Journal of Contents
• /
• 제11권1호
• /
• pp.15-20
• /
• 2015

The purpose of this research is to study methods for performing transition that have visual representation of corresponding animations with no bounce in subsequently recognized user information when attempting to interact with a virtual 3D character in real-time using user motion. If the transitions of the animation are needed owing to a variety of external environments, continuous recognition of user information is required to correspond to the motion. The proposed method includes linear interpolation of the transition using cross-fades and blending techniques. The normalized playing time of the source animation was utilized for automatically calculating the transition interpolation length of the target animation and also as the criteria in selecting the crossfades and blending techniques. In particular, in the case of blending, the weighting value based on the degree of similarity between two animations is used as a blending parameter. Accordingly, transitions for visually excellent animation are performed on interactive holographic projection systems.