Hail Song

Development of Open XR platform for high immersive collaboration

We present VRGaussianAvatar, an integrated system that enables real-time full-body 3D Gaussian Splatting (3DGS) avatars in virtual reality using only head-mounted display (HMD) tracking signals. The system adopts a parallel pipeline with a VR Frontend and a GA Backend. The VR Frontend uses inverse kinematics to estimate full-body pose and streams the resulting pose along with stereo camera parameters to the backend. The GA Backend stereoscopically renders a 3DGS avatar reconstructed from a single image. To improve stereo rendering efficiency, we introduce Binocular Batching, which jointly processes left and right eye views in a single batched pass to reduce redundant computation and support high-resolution VR displays. We evaluate VRGaussianAvatar with quantitative performance tests and a within-subject user study against image- and video-based mesh avatar baselines. Results show that VRGaussianAvatar sustains interactive VR performance and yields higher perceived appearance similarity, embodiment, and plausibility. Project page and source code are available at https://vrgaussianavatar.github.io.

We propose VR Zen Garden, a virtual reality (VR) application designed to support stress relief and meditation through interactive spatial soundscapes. The system enables users to create personalized auditory environments by combining two key interactions: Object Placement, which spatializes sound-emitting objects, and Sand Pattern Raking, which modulates sound properties through visual patterns. These inputs are processed by the Sound Spatializer and the Metaphoric Sound Mixer, allowing users to intuitively shape immersive 3D soundscapes in real time. VR Zen Garden offers an engaging and embodied approach to mindfulness, highlighting the potential of sound-driven design in immersive wellness applications.

We present a fast and efficient method for transferring facial textures onto SMPL-X–based full-body avatars. Unlike conventional affine-transform methods that are slow and prone to visual artifacts, our method utilizes a barycentric UV conversion technique. Our approach precomputes the entire UV mapping into a single transformation matrix, enabling texture transfer in a single operation. This results in a speedup of over 7000x compared to the baseline, while also significantly improving the final texture quality by eliminating boundary artifacts. Through quantitative and qualitative evaluations, we demonstrate that our method offers a practical solution for personalization in immersive XR applications. The code is available online.1

Generating high-quality 3D models from 2D sketches is a challenging task due to the inherent ambiguity and sparsity of sketch data. In this paper, we present S3D, a novel framework that converts simple hand-drawn sketches into detailed 3D models. Our method utilizes a U-Net-based encoder-decoder architecture to convert sketches into face segmentation masks, which are then used to generate a 3D representation that can be rendered from novel views. To ensure robust consistency between the sketch domain and the 3D output, we introduce a novel style-alignment loss that aligns the U-Net bottleneck features with the initial encoder outputs of the 3D generation module, significantly enhancing reconstruction fidelity. To further enhance the network's robustness, we apply augmentation techniques to the sketch dataset. This streamlined framework demonstrates the effectiveness of S3D in generating high-quality 3D models from sketch inputs. The source code for this project is publicly available at https://github.com/hailsong/S3D.

Realistic avatars are fundamental for immersive experiences in Augmented Reality (AR) and Virtual Reality (VR) environments. In this work, we introduce a novel approach for avatar generation, combining 3D Gaussian Splatting with the parametric body model, SMPL. This methodology overcomes the inefficiencies of traditional image/video-based avatar creation, which is often slow and requires high computing resources. The integration of 3D Gaussian Splatting for representing human avatar offers realistic and real-time rendering for AR/VR applications. We also conducted preliminary tests to verify the quality of avatar representation using 3D Gaussian Splatting. These tests, displayed alongside outcomes from existing methods, demonstrate the potential of this research to significantly contribute to the creation of realistic avatars in the future. Additionally, several key discussions are presented, essential for developing and evaluating the system and providing valuable insights for future research.

In avatar-mediated remote communication, avatars’ facial expressions can be dynamically adjusted according to each user’s computational and device constraints, highlighting the importance of varied expressions and their impact on user perception. However, there is a lack of research on how variations in avatar facial expressions, especially when simplified, influence user perception, particularly in terms of social presence. To address this, we examine the impact of various facial expression combinations on social presence in avatar-mediated communication scenarios, ranging from informative speeches to emotional conversations. Our approach involves prioritizing avatar facial blendshape combinations using two main approaches: (1) commonly activated expressions that reflect the active facial movements observed during casual conversations, and (2) emotion-based expressions derived from Facial Action Coding System (FACS). These combinations were compared against minimal baseline and full blendshape conditions through a comprehensive study involving 32 participants. Our findings reveal that emotion-based condition achieves comparable levels of social presence and communication quality to the full condition, in both informative speeches and emotional conversations. This highlights the effectiveness of prioritizing emotion-based expressions and adopting a streamlined approach to avatar facial control. By focusing on emotional expressions while optimizing resources, this approach shows potential for enhancing the avatar-mediated communication experience, accommodating the diverse users’ contexts.