4 papers · 1 filter
SplatReasoner: Enhancing Embodied Reasoning and Grounding by Novel View Synthesis
Kim Yu-Ji, Dahye Lee, Kim Jun-Seong +6
Vision-Language Models (VLMs) have demonstrated strong reasoning capabilities over images and videos, yet their application to embodied scene understanding often constrained by the…
SA-ResGS: Self-Augmented Residual 3D Gaussian Splatting for Next Best View Selection
Kim Jun-Seong, Tae-Hyun Oh, Eduardo Pérez-Pellitero +1
We propose Self-Augmented Residual 3D Gaussian Splatting (SA-ResGS), a novel framework to stabilize uncertainty quantification and enhancing uncertainty-aware supervision in next-b…
SoundBrush: Sound as a Brush for Visual Scene Editing
Kim Sung-Bin, Kim Jun-Seong, Junseok Ko +2
We propose SoundBrush, a model that uses sound as a brush to edit and manipulate visual scenes. We extend the generative capabilities of the Latent Diffusion Model (LDM) to incorpo…
Revisiting Learning-based Video Motion Magnification for Real-time Processing
Hyunwoo Ha, Oh Hyun-Bin, Kim Jun-Seong +6
Video motion magnification is a technique to capture and amplify subtle motion in a video that is invisible to the naked eye. The deep learning-based prior work successfully demons…