• KIPS Transactions on Software and Data Engineering
• /
• v.11 no.11
• /
• pp.465-472
• /
• 2022

In this paper, the style synthesis network is trained to generate style-synthesized video through the style synthesis through training Stylegan and the video synthesis network for video synthesis. In order to improve the point that the gaze or expression does not transfer stably, 3D face restoration technology is applied to control important features such as the pose, gaze, and expression of the head using 3D face information. In addition, by training the discriminators for the dynamics, mouth shape, image, and gaze of the Head2head network, it is possible to create a stable style synthesis video that maintains more probabilities and consistency. Using the FaceForensic dataset and the MetFace dataset, it was confirmed that the performance was increased by converting one video into another video while maintaining the consistent movement of the target face, and generating natural data through video synthesis using 3D face information from the source video's face.

• Smart Media Journal
• /
• v.10 no.1
• /
• pp.79-87
• /
• 2021

In recent years research on automatic font generation with machine learning mainly focus on using transformation-based methods, in comparison, generative model-based methods of font generation have received less attention. Transformation-based methods learn a mapping of the transformations from an existing input to a target. This makes them ambiguous because in some cases a single input reference may correspond to multiple possible outputs. In this work, we focus on font generation using the generative model-based methods which learn the buildup of the characters from noise-to-image. We propose a novel way to train a conditional generative deep neural model so that we can achieve font style control on the generated font images. Our research demonstrates how to generate new font images conditioned on both character class labels and character style labels when using the generative model-based methods. We achieve this by introducing a modified generator network which is given inputs noise, character class, and style, which help us to calculate losses separately for the character class labels and character style labels. We show that adding the character style vector on top of the character class vector separately gives the model rich information about the font and enables us to explicitly specify not only the character class but also the character style that we want the model to generate.

• The Journal of the Korea Contents Association
• /
• v.8 no.4
• /
• pp.100-107
• /
• 2008

The aim of this paper is to design an intelligent character manipulation system suitable for the wireless broadband network environment named as Wibro and for the recent social networking service (SNS). The main feature of our proposed system is an automatic synthesis of a character utilizing the meta data which is attached to each component stored in a character database. It is the second advantageous feature that a user can make or modify his or her own character by modifying the component animation clip of its character. Once the character database is created, it can be represented and utilized in several ways by utilizing the meta data which is related to personal properties including emotion, life style, favorite things, and so on. The system can be evolved to a more powerful one by adopting the more user-related information and the more intelligent synthesis module.

• Proceedings of the Korea Information Processing Society Conference
• /
• 2021.11a
• /
• pp.727-730
• /
• 2021

이미지와 비디오 합성 기술에 대한 수요가 늘어남에 따라, 인간의 손에만 의존하여 이미지나 비디오를 합성하는데에는 시간과 자원이 한정적이며, 전문적인 지식을 요한다. 이러한 문제를 해결하기 위해 최근에는 스타일 변환 네트워크를 통해 이미지를 변환하고, 믹싱하여 생성하는 알고리즘이 등장하고 있다. 이에 본 논문에서는 GAN을 이용한 스타일 변환 네트워크를 통한 자연스러운 스타일 믹싱에 대해 연구했다. 먼저 애니메이션 토이 스토리의 등장인물에 대한 데이터를 구축하고, 모델을 학습하고 두 개의 모델을 블렌딩하는 일련의 과정을 거쳐 모델을 준비한다. 그 다음에 블렌딩된 모델을 통해 타겟 이미지에 대하여 스타일 믹싱을 진행하며, 이 때 이미지 해상도와 projection 반복 값으로 스타일 변환 정도를 조절한다. 최종적으로 스타일 믹싱한 결과 이미지들을 바탕으로 하여 스타일 변형, 스타일 합성이 된 인물에 대한 동영상을 생성한다.

• Journal of the Korea Computer Graphics Society
• /
• v.26 no.4
• /
• pp.1-8
• /
• 2020

Motion synthesis using physics-based controllers can generate a character animation that interacts naturally with the given environment and other characters. Recently, various methods using deep neural networks have improved the quality of motions generated by physics-based controllers. In this paper, we present a control policy learned by deep reinforcement learning (DRL) that enables Luxo, the mascot character of Pixar animation studio, to run towards a random goal location while imitating a reference motion and maintaining its balance. Instead of directly training our DRL network to make Luxo reach a goal location, we use a reference motion that is generated to keep Luxo animation's jumping style. The reference motion is generated by linearly interpolating predetermined poses, which are defined with Luxo character's each joint angle. By applying our method, we could confirm a better Luxo policy compared to the one without any reference motions.

• The KIPS Transactions:PartB
• /
• v.12B no.7 s.103
• /
• pp.795-802
• /
• 2005

In this paper, we present a facial expression recognition-synthesis system that recognizes 7 basic emotion information automatically and renders face with non-photorelistic style in PDA For the recognition of the facial expressions, first we need to detect the face area within the image acquired from the camera. Then, a normalization procedure is applied to it for geometrical and illumination corrections. To classify a facial expression, we have found that when Gabor wavelets is combined with enhanced Fisher model the best result comes out. In our case, the out put is the 7 emotional weighting. Such weighting information transmitted to the PDA via a mobile network, is used for non-photorealistic facial expression animation. To render a 3-D avatar which has unique facial character, we adopted the cartoon-like shading method. We found that facial expression animation using emotional curves is more effective in expressing the timing of an expression comparing to the linear interpolation method.