We attempt to provide intelligence to any robots. We also aim to enable machines to understand their situation and make decisions based on their understanding.
This study addresses the challenge of performing visual localization in demanding conditions such as nighttime scenarios, adverse weather, and seasonal changes. Many prior studies tend to overlook unobserved keypoints during the matching process. Therefore, dense Read more…
Despite the prevalence of robotic manipulation tasks in various real-world applications of different requirements and needs, there has been a lack of focus on enhancing the adaptability of robotic grasping systems. Most of the current Read more…
This study addresses an image-matching problem in challenging cases, such as large scene variations or textureless scenes. To gain robustness to such situations, most previous studies have attempted to encode the global contexts of a Read more…
We propose a novel view path-planning method based on an online Multi-View Stereo (MVS) system. This method aims to incrementally construct the target three-dimensional (3D) model in real time. View paths are continually planned based Read more…
We investigate an exploration problem when constructing complete 3D models in an unknown environment using a Micro-Aerial Vehicle (MAV). Most previous exploration methods were based on the Next-Best-View (NBV) approaches, which iteratively determine the most Read more…
We investigate a path planning of a mobile robot to construct highly accurate 3D models of an unknown environment.We propose a novel exploration algorithm that considers not only a volumetric map but also reconstructed surfaces. Read more…
We instill intelligence in a humanoid robot. [KAIST AI를 선도한다]조성호 교수, 딥러닝 적용한 로봇 시각지능 기술 개발 [LINK]
We investigate the path planning problem for autonomous 3D modeling of a large outdoor structure with micro-aerial vehicle. For this problem, there are two categories of the next-best-view algorithm: volumetric and surface-based approaches. The volumetric Read more…
Navigation capability in complex and unknown outdoor environments is one of the major requirements for an autonomous vehicle and a robot that performs tasks such as a military mission or planetary exploration. Robust traversability estimation Read more…
Voxelization methods are extensively employed for efficiently processing large point clouds. However, it is possible to lose geometric information and extract inaccurate features through these voxelization methods. In this letter, we propose a novel, flexibly-shaped Read more…
We investigate a new algorithm for real-time 3D human-motion tracking through a monocular camera. The algorithm is based on body-silhouette shape matching combined with particle-filter-based selected-region tracking in the 2D view. The selected-region tracking, combined Read more…
This work proposes an algorithm for real-time navigation of a humanoid robot with a stereo vision but no other sensors. Using the algorithm, a robot can recognize its 3D environment by retrieving SIFT features from Read more…