• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.6 no.12
• /
• pp.3296-3314
• /
• 2012

We present a novel approach for the simulation of color pencil effects using line integral convolution (LIC) to produce pencil drawings from images. Our key idea is to use a bilateral convolution filter to simulate the various effects of pencil strokes. Our filter resolves the drawbacks of the existing convolution-based schemes, and presents an intuitive control to mimic the properties of pencil strokes. We also present a scheme that determines stroke directions from the shapes to be drawn. Smooth tangent flows are used for the pixels close to feature lines, and partially parallel flows inside regions. The background is rendered using a flow of fixed direction. Using different styles of stroke directions increases the realism of the resulting images. This approach produces convincing pencil drawing effects from photographs.

• Journal of Korea Multimedia Society
• /
• v.11 no.6
• /
• pp.755-763
• /
• 2008

Silhouette is useful in computer graphics for a number of techniques such as non-photorealistic rendering, silhouette clipping, and blueprint generating. Methods for generating silhouette are classified into three categories: image-based, object-based, and hybrid-based. Hybrid-based method is effective in terms of time complexity but spatial coherence problem still remains. In this paper, we proposed a new hybrid-based method which produces 3D data for silhouette and also guarantees no spatial coherence problem. To verify the efficiency of the proposed algorithm, several experiments are conducted for various 3D models from simple to quite complex. Results show that our algorithm generates no gap between any two consecutive silhouette lines when the silhouette model is magnified significantly.

• Proceedings of the Korea Information Processing Society Conference
• /
• 2009.04a
• /
• pp.208-211
• /
• 2009

일반적으로 비사실적인 렌더링(Non-Photorealistic Rendering, NPR)에 의해 렌더링된 흑백 동영상 프레임들은 재생도중에 플리커(Flicker, 깜빡거리는 현상)와 같은 부작용이 발생한다. 이것은 연속적인 프레임 영상들이 시간 일관성을 고려하지 않은 상태로 NPR 기법이 프레임 단위로 적용되기 때문이다. 그러므로 각 영상의 플리커를 완화하기 위해서는 각 프레임의 객체들이 다음 프레임에서 변형되는 정도를 측정한 후, 연속된 프레임 영상에서 객체 사이의 연관성을 설정해야 하기 때문에, NPR 기법이 적용된 연속된 영상의 디플리커는 많은 어려움을 내포하고 있다. 본 논문에서는 NPR 기법들 중, 일관성 있는 스트록을 렌더링하기 위한 NPR 기법인 Coherent Line Drawing을 적용한 이진 영상의 디플리커 방법을 제안한다.

• Proceedings of the Korean Society of Computer Information Conference
• /
• 2021.01a
• /
• pp.253-256
• /
• 2021

본 논문에서는 합성곱 신경망을 통해 학습된 DoF(피사계 심도, Depth of field) 네트워크 아키텍처를 이용하여 객체 인식, 시점 추적, 문자 인식, 비사실적 렌더링 등 다양한 애플리케이션에 적용할 수 있는 사후 필터링 기법에 대해 살펴본다. 일반적으로 영상은 포커싱과 아웃포커싱에 의해 사용자의 관심표현이 결정되며, 이를 이용하여 영상 내 중요도를 판단한다. 영상 내에는 수많은 콘텐츠들이 혼재되어 있기 때문에 사용자가 집중적으로 보고 있는 콘텐츠를 찾아내기 어렵다. 본 논문에서는 사용자가 흥미롭고 집중적으로 보고 있는 영역을 DoF 네트워크로 학습시키고, 이를 통해 이전 기법으로는 표현할 수 없었던 DoF 기반 객체 인식, 시점 추적, 문자 인식, 비사실적 렌더링을 효율적으로 표현해낸다.

• Journal of KIISE:Computer Systems and Theory
• /
• v.37 no.2
• /
• pp.110-125
• /
• 2010

We present a novel scheme for generating line drawings from 2D images, aiming to facilitate effective visual communication. In contrast to conventional edge detectors, our technique imitates the human line drawing process to generate lines effectively and intuitively. Our technique consists of three parts: line extraction, line rendering, and user guidance. In line extraction, we extract lines by estimating a likelihood function to effectively find the genuine shape boundaries. In line rendering, we consider the feature scale and the blurriness of lines with which the detail and the focus-level of lines are controlled. We also employ stroke textures to provide a variety of illustration styles. User guidance is allowed to modify the shapes and positions of lines interactively, where immediate response is provided by GPU implementation of most line extraction operations. Experimental results demonstrate that our technique generates various kinds of line drawings from 2D images enabled by the control over detail, focus, and style.

• The Journal of the Korea Contents Association
• /
• v.11 no.6
• /
• pp.106-115
• /
• 2011

In this paper, we evaluate the effectiveness of previous shading algorithms in depicting shape of 3d objects. We perform a study in which people are shown an image of one of ten 3D objects shaded with one of eight styles and asked to orient a gauge to coincide with the surface normal at many positions on the object's surface. The normal estimates are compared with each other and with ground truth data provided by a registered 3D surface model to analyze accuracy and precision. Our experiments suggest that people interpret certain shape differently depending on shading of 3d object. This paper offers substantial evidence that current computer graphics shading algorithms can effectively depict shape of 3d objects where the algorithms have the properties of lots of tone steps and uniformly distributed tone steps. This type of analysis can guide the future development of new CG shading algorithms in computer graphics for the purpose of shape perception.

• The Journal of the Korea Contents Association
• /
• v.7 no.9
• /
• pp.26-33
• /
• 2007

NPR is the area of research which does the image creation of the impression where the hazard person who creates made a human intimate image by the hand in objective. NPR in Cartoon-Rendering when the human being recognizes a thing, it is a technique which expresses a thing to use an appearance and the line which are a means of basic and intuitional expression. Also Cartoon the person probably is, the hazard which helps, raises the efficiency of namely meaning delivery and a simplification technique with the means for it uses. Is like this cartoon basic rules where the simplification and omission are life, from modeling, the low of LOD technique which it is using specially actively from game and one pulse communication is having a common point. The research which it sees Cartoon effects of GPU bases which use DirectX Shader programs LOD it leads and it applies the method for about under presenting it expresses Cartoon effects in exterior.

• Journal of the HCI Society of Korea
• /
• v.1 no.1
• /
• pp.37-44
• /
• 2006

Specular reflections provide the visual feedback that describes the material type of an object, its local shape, and lighting environment. In photorealistic rendering, there have been a number of research available to render specular reflections effectively based on a local reflection model. In traditional cel animations and cartoons, specular reflections plays important role in representing artistic intentions for an object and its related environment reflections, so the shapes of highlights are quite stylistic. In this paper, we present a method to render and control stylized specular reflections using projective textures based on principal curvature analysis. Specifying a texture as a pattern of a highlight and projecting the texture on the specular region of a given 3D model, we can obtain a stylized representation of specular reflections. For a given polygonal model, a view point, and a light source, we first find the maximum specular intensity point, and then locate the texture projector along the line parallel to the normal vector and passing through the point. The orientation of the projector is determined by the principal directions at the point. Finally, the size of the projection frustum is determined by the principal curvatures corresponding to the principal directions. The proposed method can control the position, orientation, and size of the specular reflection efficiently by translating the projector along the principal directions, rotating the projector about the normal vector, and scaling the principal curvatures, respectively. The method is be applicable to real-time applications such as cartoon style 3D games. We implement the method by Microsoft DirectX 9.0c SDK and programmable vertex/pixel shaders on Nvidia GeForce FX 7800 graphics subsystems. According to our experimental results, we can render and control the stylized specular reflections for a 3D model of several ten thousands of triangles in real-time.

• Journal of the Korea Computer Graphics Society
• /
• v.14 no.3
• /
• pp.1-9
• /
• 2008

This paper presents a novel optimization-based saliency-preserving method for converting color images to grayscale in a manner consistent with conventional approaches of black-and-white photographers. In black-and-white photography, a colored filter called a contrast filter has been commonly employed on a camera to lighten or darken selected colors. In addition, local exposure controls such as dodging and burning techniques are typically employed in the darkroom process to change the exposure of local areas within the print without affecting the overall exposure. Our method seeks a digital version of a conventional contrast filter to preserve visually-important image features. Furthermore, conventional burning and dodging techniques are addressed, together with image similarity weights, to give edge-aware local exposure control over the image space. Our method can be efficiently optimized on GPU. According to the experiments, CUDA implementation enables 1 megapixel color images to be converted to grayscale at interactive frames rates.

• Journal of the Korea Society of Computer and Information
• /
• v.26 no.3
• /
• pp.51-57
• /
• 2021

In this paper, we propose a method to find the DoF(Depth of field) that is blurred in an image by focusing and out-focusing the camera through a efficient convolutional neural network. Our approach uses the RGB channel-based cross-correlation filter to efficiently classify the DoF region from the image and build data for learning in the convolutional neural network. A data pair of the training data is established between the image and the DoF weighted map. Data used for learning uses DoF weight maps extracted by cross-correlation filters, and uses the result of applying the smoothing process to increase the convergence rate in the network learning stage. The DoF weighted image obtained as the test result stably finds the DoF region in the input image. As a result, the proposed method can be used in various places such as NPR(Non-photorealistic rendering) rendering and object detection by using the DoF area as the user's ROI(Region of interest).