이미지 기반 모델링을 활용한 3D프린팅에서 디스플레이스먼트맵에 의한 표면질감 향상

김진곤 사단법인 한국브랜드디자인학회 2022 브랜드디자인학연구 Vol.20 No.3

Recently, image-based modeling and 3D printing are being actively attempted in various product development environments. However, despite the usefulness of the technology, image-based modeling may cause problems of different quality from the original in terms of surface texture implementation. The purpose of this study was to explore ways to effectively compensate for the loss of surface texture in a workflow that processes polygon mesh data created using image-based modeling into 3D printing to improve the quality of the final output. To this end, image-based modeling technology was applied to several experiments, converted into polygon mesh, and the degree of reproduction of the original surface such as texture and flexion was examined, and a diffuser map was converted into a display map to displace its phase. In the display map, the result of phase displacement was identified by deepening the depth through contrast adjustment of black corresponding to the lowest point of height value and white corresponding to the highest point. As a result, it was possible to obtain a result that the surface quality was improved due to the application of the display map with controlled light and dark gradation for low-precision and unclear curves. Finally, the quality of the final polygon mesh data with improved surface flexion, pattern detail, and reproducibility was confirmed after 3D printing, and the surface texture was clearer than the polygon mesh quality derived through image-based modeling. 근래 다양한 상품개발 환경에서 이미지 기반 모델링과 3D프린팅의 활용이 적극적으로 시도되고 있는 추세다. 그런데 이미지 기반 모델링은 그 기술의 유용함에도 불구하고 표면질감 구현측면에서 원본과 다른 품질의 문제가 발생하기도 한다. 따라서 본 연구는 이미지 기반 모델링을 이용해 만들어진 폴리곤 메쉬 데이터를 3D프린팅으로 가공하는 워크플로우에서 표면질감의 손실을 효과적으로 보완하여 최종 출력물의 품질을 향상시킬 수 있는 방법에 대한 탐색을 목적으로 하였다. 이를 위해 몇 가지 실험물을 대상으로 이미지 기반 모델링 기술을 적용하여 폴리곤 메쉬로 변환한 후 질감, 굴곡 등 원본 표면의 재현 정도를 살피고, 디퓨즈맵을 디스플레이스먼트맵으로 변환하여 위상을 변위시키는 실험을 진행하였다. 디스플레이스먼트맵에서 높이값의 최저점에 해당하는 검정색 및 최고점에 해당하는 흰색의 콘트라스트 조절을 통해 깊이감을 심화시킴으로써 위상 변위의 결과를 파악해본 것이다. 그 결과 세밀함 정도가 낮고 뚜렷하지 않던 굴곡이 명암 계조가 조절된 디스플레이스먼트맵의 적용으로 인해 표면품질이 향상되는 결과를 얻을 수 있었다. 최종적으로 디스플레이스먼트맵이 적용되어 표면굴곡이나 패턴의 세밀함, 재현력이 향상된 최종 폴리곤 메쉬 데이터를 3D프린팅 출력 후 품질을 확인하였는데, 처음 이미지 기반 모델링을 통해 도출된 폴리곤 메쉬 품질보다 더 선명한 표면질감을 얻을 수 있었다.

물체 표면에 광택을 가진 이미지를 활용한 3D 재구성 연구

장가오허,김기홍,후정동,리펑휘 사단법인 한국융합기술연구학회 2023 아시아태평양융합연구교류논문지 Vol.9 No.1

Image-based modeling and rendering technology is characterized by its low cost, high efficiency, and simple operation compared to other modeling methods, but it still has certain limitations. Image-based modeling and rendering techniques, also known as image modeling, are the task of taking tens or hundreds of photographs and then reconstructing a 3D model of a subject with a related software processing system using mathematical methods. However, during the reconstruction process, if a portion of the object is too glossy on the surface, a mirror reflection phenomenon will exist. When a digital camera photographs the surface of a high-gloss object obscured by mirror reflection, the original color and texture information of the digital image is lost, which may result in a deviation in the model in reconstruction operations that image.Thus, addressing the effects of glossiness on information on the surface of objects in digital images in an effective manner has significance in the field of image modeling. This paper studied the principle of image modeling technology. After measuring the glossiness of an object using a gloss meter, a reconstruction experiment was conducted to confirm the effect of the glossiness of the object on the image modeling technology. For image modeling, the data were referenced and various methods of removing gloss information were compared through experiments.We also present a method for restoring and identifying reflected light points (highlights) on the surface of a high-gloss object using the gray scale information of the image. This method can significantly improve the reconstruction effect of high-gloss objects in certain environments, and the results of the experiment show that the reconstructed image after the highlight was restored increased by 14% compared to the pre-restored image.

A model-based radiography restoration method based on simple scatter-degradation scheme for improving image visibility

Kim, K.,Kang, S.,Cho, H.,Kang, W.,Seo, C.,Park, C.,Lee, D.,Lim, H.,Lee, H.,Kim, G.,Park, S.,Park, J.,Kim, W.,Jeon, D.,Woo, T.,Oh, J. Elsevier 2018 Optics and lasers in engineering Vol.101 No.-

<P><B>Abstract</B></P> <P>In conventional planar radiography, image visibility is often limited mainly due to the superimposition of the object structure under investigation and the artifacts caused by scattered x-rays and noise. Several methods, including computed tomography (CT) as a multiplanar imaging modality, air-gap and grid techniques for the reduction of scatters, phase-contrast imaging as another image-contrast modality, <I>etc.</I>, have extensively been investigated in attempt to overcome these difficulties. However, those methods typically require higher x-ray doses or special equipment. In this work, as another approach, we propose a new model-based radiography restoration method based on simple scatter-degradation scheme where the intensity of scattered x-rays and the transmission function of a given object are estimated from a single x-ray image to restore the original degraded image. We implemented the proposed algorithm and performed an experiment to demonstrate its viability. Our results indicate that the degradation of image characteristics by scattered x-rays and noise was effectively recovered by using the proposed method, which improves the image visibility in radiography considerably.</P> <P><B>Highlights</B></P> <P> <UL> <LI> It is proposed for a model-based radiography restoration method. </LI> <LI> The intensity of scattered x-rays and the transmission function are estimated. </LI> <LI> We implement the proposed algorithm. </LI> <LI> This study improves the image visibility in radiography considerably. </LI> </UL> </P>

항공사진과 수치지도를 이용한 도시 건물의 이미지 기반 모델링

유병현,한순흥 한국지리정보학회 2005 한국지리정보학회지 Vol.8 No.1

비행 시뮬레이션과 같은 가상현실 시스템은 실제 존재하는 도시의 가상환경을 필요로 하며, 항상 변화하고 있는 도시환경과 치대한 일치하는 도시모델을 유지하여야 한다. 따라서 주기적으로 갱신되는 지형공간정보를 활용할 수 있는 도시 모델링 방법이 필요하다. 또한 비행 시뮬레이션은 가까운 지역뿐만이 아니라 먼 거리의 건물을 효과적으로 표현할 수 있어야 한다. 이를 위하여 본 논문에서는 고해상도 항공사진과 수지지도로부터, 가상 도시환경을 비교적 적은 노력으로 모델링할 수 있는 방법을 제시한다. 이미지 운반의 모델링 기법을 아파트처럼 단순한 구조의 건물이 많은 한국의 도시환경과 지형공간정보의 특성에 맞도록 적용하여, 수많은 건물을 폴리곤으로 모델링하지 않고 효율적으로 표현할 수 있도록 한다. 세안하는 방법은 지형공간정보의 전처리 과정과 가상 도시환경 모델링 과정으로 구성되며, 건물의 높이 맵을 만드는 간단한 전처리 과정만으로 가상도시환경을 모델링 할 수 있어, 기존의 폴리곤 기반의 모델링 방법이 대규모 도시 가상환경에서 멀리 있는 수많은 건물의 표현을 생략하였던 문제점을 개선하였다. The VR (virtual reality) simulator such as helicopter simulation needs virtual environment of existing urban area. But the real urban environment keeps changing. We need a modeling method to make use of the GIS data that are updated periodically. The flight simulation needs to visualize not only buildings in near distance but also a large number of buildings in the far distance. We propose a method for modeling urban environment from aerial image and digital map with a comparatively small manual work. Image based modeling is applied to urban model which considers the characteristic of Korean cities. Buildings in the distance can be presented without creating a lot of polygons. Proposed method consists of the pre-processing stage which prepares the model from the GIS data and the modeling stage which makes the virtual urban environment. The virtual urban environment can be modeled with the simple process which utilizes the height map of buildings.

영상 관찰 모델을 이용한 예제기반 초해상도 텍스트 영상 복원

김인중,박규로 한국정보처리학회 2010 정보처리학회논문지. 소프트웨어 및 데이터 공학 Vol.17 No.4

Example-based super resolution(EBSR) is a method to reconstruct high-resolution images by learning patch-wise correspondence between high-resolution and low-resolution images. It can reconstruct a high-resolution from just a single low-resolution image. However, when it is applied to a text image whose font type and size are different from those of training images, it often produces lots of noise. The primary reason is that, in the patch matching step of the reconstruction process, input patches can be inappropriately matched to the high-resolution patches in the patch dictionary. In this paper, we propose a new patch matching method to overcome this problem. Using an image observation model, it preserves the correlation between the input and the output images. Therefore, it effectively suppresses spurious noise caused by inappropriately matched patches. This does not only improve the quality of the output image but also allows the system to use a huge dictionary containing a variety of font types and sizes, which significantly improves the adaptability to variation in font type and size. In experiments, the proposed method outperformed conventional methods in reconstruction of multi-font and multi-size images. Moreover, it improved recognition performance from 88.58% to 93.54%, which confirms the practical effect of the proposed method on recognition performance. 예제기반 초해상도 영상 복원(EBSR)은 고해상도 영상과 저해상도 영상간의 패치간 대응관계를 학습함으로써 고해상도 영상을 복원하는 방법으로, 한 장의 저해상도 영상으로부터도 고해상도 영상을 복원할 수 있는 장점이 있다. 그러나, 폰트의 종류나 크기가 학습 영상과 다른 텍스트 영상을 적용할 경우 잡영을 많이 발생시킨다. 그 이유는 복원 과정 중 매칭 단계에서 입력 패치들이 사전 내의 고해상도 패치와 부적절하게 매칭될 수 있기 때문이다. 본 논문에서는 이러한 문제점을 극복하기 위한 새로운 패치 매칭 방법을 제안한다. 제안하는 방법은 영상 관찰 모델을 이용하여 입력 영상과 출력 영상간의 상관 관계를 보존함으로써 잘못 매칭된 패치로 인한 잡영을 효과적으로 억제한다. 이는 출력 영상의 화질을 개선할 뿐 아니라, 다양한 종류 및 크기의 폰트를 포함한 대용량 패치 사전을 적용할 수 있게 함으로써 폰트의 종류 및 크기의 변이에 대한 적응력을 크게 향상시킨다. 실험에서 제안하는 방법은 폰트와 크기가 다양한 영상에 대하여 기존의 방법보다 우수한 영상 복원 성능을 나타내었다. 뿐만 아니라, 인식 성능도 88.58%에서 93.54%로 개선되어 제안하는 방법이 인식 성능의 개선에도 효과적임을 확인하였다.

DIGITAL 3D IMAGE COMPOSITING USING IMAGE-BASED MODELING

Chang Ok Yun,Young Bo Lee,Tae Soo Yun,Dong Hoon Lee 한국멀티미디어학회 2006 한국멀티미디어학회 국제학술대회 Vol.2006 No.-

Due to recent advances in digital technology, people are showing an increased interest in film and video technology. Image compositing is the core of the digital image related multimedia technology. Currently, various digital image compositing technologies are being developed. So far, digital compositing techniques make use of two general methods. The first is compositing technique of using motion control cameras for precisely capturing in 3D camera motion information. The second is a compositing technique of combining 3D modeling by using 3D graphics editing tool with the existing digital source images. However, when the compositing object is 2D photograph, digital compositing has a great deal of difficulty combining with digital source images properly. In this paper, in order to overcome common difficulties, we propose a new digital 3D image compositing method aimed to generate 3D space model for modeling scenes which used estimated geometric information and combine with digital source images. At this time, we generate a 3D space model by using image-based modeling from a general single 2D photograph. Therefore, we can generate compositing results of high quality easily and quickly with a 2D photograph. Moreover, it is possible for us to generate a more efficient and actual compositing image.

Variational hand modeling based on image-based 3D model reconstruction

Yulai Xie,Satoshi Kanai,Hiroaki Date (사)한국CDE학회 2013 한국CAD/CAM학회 국제학술발표 논문집 Vol.2010 No.8

A method for variational modeling of individual hands from a template hand model was described. The marked individual hand model was reconstructed based on an image-based multiview 3D model reconstruction technique. The skin surface of the template hand model was partitioned into segments whose boundaries were taken as main creases on palmar side according to the hand surface anatomy. The segments of the template hand model are scaled according to the feature dimensions of the segments of the reconstructed individual hand model. On the other hand, according to hand joint anatomical knowledge, a single axis rotation assumption was proposed for estimating the joint axes and joint centers by the positions of dot markers at the backside of the hand skin in different hand postures. The joint axes and their centers could be estimate from the correspondences of the dot marker positions painted at the back side of the hand in different postures. Finally, the scaled template hand model was in a segment-wise way was registered to the reconstructed hand model by an alignment process. Compared with previous works, the method was more subject-friendly, inexpensive and efficient.

City-Scale Modeling for Street Navigation

Huang, Fay,Klette, Reinhard The Korea Institute of Information and Commucation 2012 Journal of information and communication convergen Vol.10 No.4

This paper proposes a semi-automatic image-based approach for 3-dimensional (3D) modeling of buildings along streets. Image-based urban 3D modeling techniques are typically based on the use of aerial and ground-level images. The aerial image of the relevant area is extracted from publically available sources in Google Maps by stitching together different patches of the map. Panoramic images are common for ground-level recording because they have advantages for 3D modeling. A panoramic video recorder is used in the proposed approach for recording sequences of ground-level spherical panoramic images. The proposed approach has two advantages. First, detected camera trajectories are more accurate and stable (compared to methods using multi-view planar images only) due to the use of spherical panoramic images. Second, we extract the texture of a facade of a building from a single panoramic image. Thus, there is no need to deal with color blending problems that typically occur when using overlapping textures.

시계열 이미지 데이터 기반 상품추천을 위한 CNN 모델 성능 비교 연구

심준현,김철진 한국지식정보기술학회 2023 한국지식정보기술학회 논문지 Vol.18 No.5

In the modern world, advances in information technology have led to the expansion of e-commerce, making it important for automated recommendation systems to efficiently gather the flood of information and data to present consumers with their favorite products and services. Various techniques are used to improve the accuracy of product recommendation in existing e-commerce. Among them, there are chronic problems that use RNN, a multi classification-based product recommendation model. RNN is a deep learning model suitable for time series classification tasks, but it suffers from issues such as gradient vanishing and gradient exploding. To Compensate for these issues, CNN models are often used to effectively detect local patterns through kernels. In this study, we compare the performance of recommendation models based on an architecture that generates product recommendation models by training CNN models with time series data through three different imaging encodings: GAF, MTF and RP. In our experiments, we split the 540,000 published transaction dataset into train and test. The splitted data is constructed as time series data and zero-padded to equalize the size of the model’s input image. We train AlexNet, VGG16, ResNet50, and MobileNet models on images generated by the three imaging algorithms and compare their product recommendation accuracy with the performance of existing RNN recommendation models. We can see that the CNN models perform better than the LSTM. When imaged with the GAF algorithm and trained on the MobileNet model, the highest recommendation accuracy was achieved, and the learning time was also shortened, improving efficiency. Future research will include the advancement of imaging algorithms to improve the performance of product recommendation models and the development of CNN models optimized for time series image data.

복부 CT에서 Adaptive Statistical Iterative Reconstruction과 Model-Based Iterative Reconstruction의 사용에 따른 선량 감소 및 화질 평가

손은영(Eun young Son),홍선숙(Sun suk Hong) 대한CT영상기술학회 2013 대한CT영상기술학회지 Vol.15 No.1

목적 본 연구는 복부 CT 검사에서 adaptive statistical iterative reconstruction 기법과 model based iterative reconstruction 기법으로 획득한 이미지를 토대로 복부 장기의 effective dose 변화 및 화질의 정량적인 분석을 통하여 최적의 영상 정보를 제공할 수 있는 iterative reconstruction을 알아보고자 한다. 대상 및 방법 본 연구를 위해 CT 장비는 GE사의 64-slice multidetector-row CT Discovery 750 HD를 사용하였고, 인체 모형 Rando phantom을 사용하였으며, 복부 장기 effective dose 측정을 위하여 유리 선량계를 사용하였다. 장비의 성능 평가를 하기 위하여 ACR phantom을 사용하였고, Image J와 Matlab program을 사용하여 noise, SNR, distortion 및 resolution 측정을 통해 화질을 정량적으로 평가, 분석하였다. 결과 DLP 및 effective dose는 인위적으로 선량을 감소하여 ASIR와 MBIR로 재구성함에 따라 그 비율대로 유사하게 선량이 감소하였다. Noise는 120 mAs-ASIR(ASIR 40%)에서 최소, 100 mAs-ASIR(ASIR 50%)부터 증가하는 것으로 나타났으며, MBIR은 dose를 감소하지 않고 reconstruction하였을 경우 최소로 측정되어 dose가 감소함에 따라 noise는 증가하는 것으로 나타났다. SNR의 경우 180 mAs-ASIR(ASIR 10%)와 140 mAs-MBIR에서 가장 높게 나타났다. 현재 복부 CT에 사용되는 조건의 영상을 기준으로 image의 distortion을 평가했을 때 80 mAs-ASIR(ASIR 70%)까지 PSNR이 30 dB을 넘게 측정되었으며, MBIR은 모두 30 dB 이하로 측정되었다. 해상력과 선예도 측정 결과에서는 100 mAs-MBIR이 100 mAs-ASIR(ASIR 50%)에 비해 sharpness는 20.58%, resolution은 33.19% 높게 측정되었다. ACR phantom을 통한 장비의 성능 평가는 모두 적합한 것으로 나타났다. 결론 본 연구는 복부 CT 검사에서 200 mAs부터 60 mAs까지 10%씩 선량을 감소하여 ASIR와 MBIR로 재구성한 이미지를 토대로 복부 장기의 effective dose 변화 및 화질을 평가하기 위하여 진행하였다. 결과 모두 환자의 선량을 감소시켜 ASIR와 MBIR로 재구성하여도 진단 가치가 있는 것을 확인할 수 있었으며, 이러한 기법을 복부 CT protocol에 적용한 결과 180 mAs-ASIR(ASIR 10%)에서 최적화된 정보를 얻을 수 있었다. MBIR의 경우 ASIR에 비하여 같은 선량에서 약 67%의 많은 noise를 감소하며, 해상력이 높은 것으로 나타났으나 현재 실용화하기 위해서는 고성능의 컴퓨터와 많은 재구성 시간을 요구하는 단점을 가지고 있다. Ⅰ. Purpose The purpose of this study was to assess iterative reconstruction methods that provide optimal image information in abdominal CT exams. It was based on the acquired images by adaptive statistical iterative reconstruction (ASIR) and model- based iterative reconstruction (MBIR). We analyzed the change of effective dose in abdominal organs and image quality quantitatively. Ⅱ. Meterial and Methods CT scans were performed with a 64-slice multidetector-row CT Discovery 750HD and Rando phantom. Glass dosimeters were used to measure effective dose of organs. ACR phantom was used for performance evaluation of the equipment. We quantitatively evaluated and analyzed image quality by measurement of noise, SNR, distortion, and resolution by using Image J and Matlab program. Ⅲ. Result As we gradually lowered radiation dose and reconstructed with ASIR and MBIR, the value of DLP and effective dose also decreased similarly. Noise value was the lowest at 120 mAs-ASIR (ASIR 40%), and got increased from 100 mAs-ASIR (ASIR 50%). SNR value was the highest at 180 mAs-ASIR (ASIR 10%) and 140 mAs-MBIR. In evaluation of distortion of image quality based on abdominal CT protocol, PSNR showed over 30 dB up to 80 mAs-ASIR (ASIR 60%), and below 30 dB at all MBIRs. In measurement of resolution, sharpness and resolving power at 100 mAs-MBIR were respectively 20.58% and 33.19% higher than those at 180 mAs-ASIR (ASIR 10%). All performance evaluation of the equipment through ACR phantom was found to be acceptable. Ⅳ. Conclusions There’s diagnostic value with both iterative reconstruction methods, ASIR and MBIR, while reducing patient dose. In abdominal CT protocol, we can get optimal information at 180 mAs-ASIR (ASIR 10%). MBIR can provide 67% less noise and higher resolution compared to ASIR in the same dose, however, it requires high performance computer and considerable time to reconstruct.