• Journal of Broadcast Engineering
• /
• v.4 no.2
• /
• pp.87-102
• /
• 1999

We propose an algorithm for augmenting a real video sequence with views of graphics ojbects without metric calibration of the video camera by representing the motion of the video camera in projective space. We define a virtual camera, through which views of graphics objects are generated. attached to the real camera by specifying image locations of the world coordinate system of the virtual world. The virtual camera is decomposed into calibration and motion components in order to make full use of graphics tools. The projective motion of the real camera recovered from image matches has a function of transferring the virtual camera and makes the virtual camera move according to the motion of the real camera. The virtual camera also follows the change of the internal parameters of the real camera. This paper shows theoretical and experimental results of our application of non-metric vision to augmented reality.

• Journal of the Korea Computer Graphics Society
• /
• v.27 no.5
• /
• pp.33-43
• /
• 2021

In this paper, we propose a framework that transfers the information of style motions to content motions based on a variational autoencoder network combined with a style encoding in the latent space. Because we transfer a style to a content motion that is sampled from a variational autoencoder, we can increase the diversity of existing motion data. In addition, we can improve the unnatural motions caused by decoding a new latent variable from style transfer. That improvement was achieved by additionally using the velocity information of motions when generating next frames.

• Journal of the Korea Computer Graphics Society
• /
• v.27 no.5
• /
• pp.73-79
• /
• 2021

Recently, in the field of computer animation, a method for generating motion using deep learning has been studied away from conventional finite-state machines or graph-based methods. The expressiveness of the network required for learning motions is more influenced by the diversity of motion contained in it than by the simple length of motion to be learned. This study aims to find an efficient network structure when the types of motions to be learned are diverse. In this paper, we train and compare three types of networks: basic fully-connected structure, mixture of experts structure that uses multiple fully-connected layers in parallel, recurrent neural network which is widely used to deal with seq2seq, and transformer structure used for sequence-type data processing in the natural language processing field.

• Journal of the Korea Computer Graphics Society
• /
• v.29 no.3
• /
• pp.59-67
• /
• 2023

This paper proposes a 3D animated body profile using 3D body scanning and motion capture devices. Users can create their own personalized body profiles with animation by performing 3D scans for a predetermined set of poses. To achieve this, a template animation was obtained through motion capture for a series of poses, and the acquired 3D scan data from users was mapped to the key poses of the animation using Pose-space deformer. The resulting 3D animated body profiles provide users with greater satisfaction compared to traditional static 2D images or 3D scan data.

• Journal of the Korea Computer Graphics Society
• /
• v.5 no.2
• /
• pp.33-41
• /
• 1999

This paper presents a new algorithm about motion adjustment for dynamic balance. It adjusts an unbalanced motion to an balanced motion while preserving the nuance of original motion. We solve dynamic balancing problem using the zero moment point (ZMP) which is often used for controlling the balance of biped robot. Our algorithm is consists of four steps. First, it fits joint angle data to spline curves for reducing noise. Second, the algorithm analyzes the ZMP trajectory so that it can detects the dynamically-unbalanced duration. Third, the algorithm project the ZMP trajectory into the supporting area if the trajectory deviates from the area. Finally, the algorithm produces the balanced motion that satisfies the new ZMP trajectory. In this step, the constrained optimization method is used so that the new motion keeps the original motion characteristics as much as possible. We make several experiments in order to prove that our algorithm is useful to add physical realism to a kinematically edited motion.

• Journal of the Korea Computer Graphics Society
• /
• v.21 no.5
• /
• pp.29-36
• /
• 2015

Human fingers are essential parts of the body that perform complex and detailed motion. Expression of natural finger motion is one of the most important issues in character animation research. Especially, keyboard typing animation is hard to create through the existing animation pipeline because the keyboard typing typically requires a high level of dexterous motion that involves the movement of various joints in a natural way. In this paper, we suggest a method for the generation of realistic keyboard typing motion based on physics simulation. To generate typing motion properly using physics-based simulation, the hand and the keyboard models should be positioned in an allowed range of simulation space, and the typing has to occur at a precise key location according to the input signal. Based on the observation, we incorporate natural tendency that accompanies actual keyboard typing. For example, we found out that the positions of the hands and fingers always assume the default pose, and the idle fingers tend to minimize their motion. We handle these various constraints in one solver to achieve the results of real-time natural keyboard typing simulation. These results can be employed in various animation and virtual reality applications.

• Journal of the Korea Computer Graphics Society
• /
• v.15 no.4
• /
• pp.31-40
• /
• 2009

Captured motion data is widely used today in a variety of areas including film production, game development, sports analysis, and medical rehabilitation. The ability of analyzing and processing motion data has increased rapidly for last decades. However, it is still difficult for users to quickly understand the contents of motion data consisting of a series of time-varying poses. One typical approach is to visualize consecutive poses in sequence while adjusting three-dimensional view, which is often time-consuming and laborious especially when users need to repeatedly control time and view in order to search for desired motions. We present a method of visualizing motion data as a sequence of images in comic book layout so that users can rapidly understand the overall flows of motion data, and easily identify their desired motions. The usefulness of our approach is demonstrated by visualizing various kinds of motion data including locomotion, boxing, and interaction with environments.

• Journal of the Korea Computer Graphics Society
• /
• v.13 no.3
• /
• pp.1-6
• /
• 2007

Anticipation effect has been used as a traditional skill to enhance the dynamic motion of the traditional 2D animation. Basically, anticipation means the action of opposite direction which performs before the real action step. In this paper, we propose the perception-based video anticipation method to guide a user's visual attention to the important region. Using the image based attention map, we calculate the visual attention region and then combine this map with temporal saliency of video. We apply the anticipation effect in these saliency regions using the blur kernel. Using our method, we can generate the dynamic video motion which has attentive guidance.

• Journal of the Korea Computer Graphics Society
• /
• v.27 no.1
• /
• pp.1-8
• /
• 2021

This paper proposes a simple deformation method for editing the trajectory of a walking motion with preserving its style. To this end, our method analyzes the trajectory of the root joint into the graph and deforms it by applying the graph Laplace operator. The trajectory of the root joint is presented as a graph with a vertex defined the position and direction at each time frame on the motion dataThe graph transforms the trajectory into the differential coordinate, and if the constraints are set on the trajectory vertex, the solver iterative approaches to the solution. By modifying the root trajectory, we can continuously vary the walking motion, which reduces the cost of capturing a whole motion that is required. After computes the root trajectory, other joints are copied on the root and post-processed as a final motion. At the end of our paper, we show the application that the character continuously walks in a complex environment while satisfying user constraints.

• Journal of the Korea Computer Graphics Society
• /
• v.5 no.2
• /
• pp.25-32
• /
• 1999

본 논문은 한 캐릭터의 동작을 다른 캐릭터에게도 이용 가능하도록 실시간으로 동작을 변환하는 알고리즘을 제시한다. 본 알고리즘은 작업 우선 순위를 고려한 폐루프 역 변화율 제어(closed-loop inverse rate control)에 기반하고 있다. 최우선 순위의 작업으로서 캐릭터간의 앤드이펙터들의 궤적의 차이를 줄이도록 하고, 다음 우선 순위의 작업으로 잉여 자유도를 이용하여 캐릭터간의 관절각의 차이를 최소화함으로서 전체 동작 변환이 수행된다. 동작 변환은 온라인으로 이루어지므로 모션 캡쳐시 변환되는 동작을 화면상에서 실시간으로 볼 수 있다. 따라서 동작을 수행하는 사람이 원하는 결과가 얻어질 때까지 화면을 보면서 동작을 변화시킬 수 있으며 이는 오프라인 알고리즘에 비해 보다 효과적인 인터액션을 가능하게 한다.