12 papers
RAIN: Robotic Task Generalization with Region-Aware Interaction Networks
Seunghun Lee, Kyuhwan Shim, Yeogyeong Song +5
Existing vision-language-action (VLA) models benefit from large-scale vision-language pretraining, but tend to memorize training trajectories, limiting their ability to adapt learn…
CANTABILE: Learning Expressive Dynamics for Robotic Piano Performance
Woosik Kim, Wonhyeok Choi, Sunghoon Im
Robotic piano playing has emerged as a standard benchmark for dexterous bimanual manipulation, yet progress on it has been measured almost entirely by note accuracy -- which keys a…
CRAFT: Constrained Reward via Attention Fine-Tuning for Subject Personalization without Composed Targets
Jihun Park, Kyoungmin Lee, Jongmin Gim +6
Subject-driven image personalization---generating new images that preserve the identity of one or several reference subjects in novel scenes---is a foundational capability for mode…
CascadeOcc: Rethinking 3D Occupancy World Models with Cascaded VQ Representations
Kyumin Hwang, Wonhyeok Choi, Jaeyeul Kim +3
This letter proposes CascadeOcc, a novel occupancy world model that prioritizes intrinsic structural hierarchy over extrinsic auxiliary modalities for autonomous driving. Occupancy…
A Review of Online Diffusion Policy RL Algorithms for Scalable Robotic Control
Wonhyeok Choi, Shutong Ding, Minwoo Choi +5
Diffusion policies have emerged as a powerful approach for robotic control, demonstrating superior expressiveness in modeling multimodal action distributions compared to convention…
Scale-invariant and View-relational Representation Learning for Full Surround Monocular Depth
Kyumin Hwang, Wonhyeok Choi, Kiljoon Han +5
Recent foundation models demonstrate strong generalization capabilities in monocular depth estimation. However, directly applying these models to Full Surround Monocular Depth Esti…
Infinite-Story: A Training-Free Consistent Text-to-Image Generation
Jihun Park, Kyoungmin Lee, Jongmin Gim +7
We present Infinite-Story, a training-free framework for consistent text-to-image (T2I) generation tailored for multi-prompt storytelling scenarios. Built upon a scale-wise autoreg…
A Training-Free Style-Personalization via SVD-Based Feature Decomposition
Kyoungmin Lee, Jihun Park, Jongmin Gim +4
We present a training-free framework for style-personalized image generation that operates during inference using a scale-wise autoregressive model. Our method generates a stylized…
Bridging Geometric and Semantic Foundation Models for Generalized Monocular Depth Estimation
Sanggyun Ma, Wonjoon Choi, Jihun Park +4
We present Bridging Geometric and Semantic (BriGeS), an effective method that fuses geometric and semantic information within foundation models to enhance Monocular Depth Estimatio…
A Training-Free Style-aligned Image Generation with Scale-wise Autoregressive Model
Jihun Park, Jongmin Gim, Kyoungmin Lee +5
We present a training-free style-aligned image generation method that leverages a scale-wise autoregressive model. While large-scale text-to-image (T2I) models, particularly diffus…
Flow4D: Leveraging 4D Voxel Network for LiDAR Scene Flow Estimation
Jaeyeul Kim, Jungwan Woo, Ukcheol Shin +2
Understanding the motion states of the surrounding environment is critical for safe autonomous driving. These motion states can be accurately derived from scene flow, which capture…
Rethinking LiDAR Domain Generalization: Single Source as Multiple Density Domains
Jaeyeul Kim, Jungwan Woo, Jeonghoon Kim +1
In the realm of LiDAR-based perception, significant strides have been made, yet domain generalization remains a substantial challenge. The performance often deteriorates when model…