arXiv subjects

Takuya Ikeda

Publications and source records attributed to Takuya Ikeda.

At least 19 recordsLinked to original sources

arXiv2025

OmniShape: Zero-Shot Multi-Hypothesis Shape and Pose Estimation in the Real World

We would like to estimate the pose and full shape of an object from a single observation, without assuming known 3D model or category. In this work, we propose OmniShape, the first method of its kind to enable probabilistic pose and shape estimation. OmniShape is based on the key insight that shape completion can be decoupled into two multi-modal distributions: one capturing how measurements project into a normalized object reference frame defined by the dataset and the other modelling a prior over object geometries represented as triplanar neural fields. By training separate conditional diffusion models for these two distributions, we enable sampling multiple hypotheses from the joint pose and shape distribution. OmniShape demonstrates compelling performance on challenging real world datasets. Project website: https://tri-ml.github.io/omnishape

cs.CV↗

arXiv2025

Sparse optimal control for infinite-dimensional linear systems with applications to graphon control

Large-scale networked systems typically operate under resource constraints, and it is also difficult to exactly obtain the network structure between nodes. To address these issues, this paper investigates a sparse optimal control for infinite-dimensional linear systems and its application to networked systems where the network structure is represented by a limit function called a graphon that captures the overall connection pattern. The contributions of this paper are twofold: (i) To reduce computational complexity, we derive a sufficient condition under which the sparse optimal control can be obtained by solving its corresponding L1 optimization problem. Furthermore, we introduce a class of non-convex optimal control problems such that the optimal solution always coincides with a sparse optimal control, provided that the non-convex problems admit optimal solutions. (ii) We show that the sparse optimal control for large-scale finite-dimensional networked systems can be approximated by that of the corresponding limit graphon system, provided that the underlying graph is close to the limit graphon in the cut-norm topology. The effectiveness of the proposed approach is illustrated through numerical examples.

math.OC↗

arXiv2025

GTR: Gaussian Splatting Tracking and Reconstruction of Unknown Objects Based on Appearance and Geometric Complexity

We present a novel method for 6-DoF object tracking and high-quality 3D reconstruction from monocular RGBD video. Existing methods, while achieving impressive results, often struggle with complex objects, particularly those exhibiting symmetry, intricate geometry or complex appearance. To bridge these gaps, we introduce an adaptive method that combines 3D Gaussian Splatting, hybrid geometry/appearance tracking, and key frame selection to achieve robust tracking and accurate reconstructions across a diverse range of objects. Additionally, we present a benchmark covering these challenging object classes, providing high-quality annotations for evaluating both tracking and reconstruction performance. Our approach demonstrates strong capabilities in recovering high-fidelity object meshes, setting a new standard for single-sensor 3D reconstruction in open-world environments.

cs.CV↗

arXiv2025

Switched Systems Control via Discreteness-Promoting Regularization

This paper proposes a novel method for designing finite-horizon discrete-valued switching signals in linear switched systems based on discreteness-promoting regularization. The inherent combinatorial optimization problem is reformulated as a continuous optimization problem with a non-convex regularization term that promotes discreteness of the control. We prove that any solution obtained from the relaxed problem is also a solution to the original problem. The resulting non-convex optimization problem is efficiently solved through time discretization. Numerical examples demonstrate the effectiveness of the proposed method.

math.OC↗

arXiv2025

Simultaneous Pick and Place Detection by Combining SE(3) Diffusion Models with Differential Kinematics

Grasp detection methods typically target the detection of a set of free-floating hand poses that can grasp the object. However, not all of the detected grasp poses are executable due to physical constraints. Even though it is straightforward to filter invalid grasp poses in the post-process, such a two-staged approach is computationally inefficient, especially when the constraint is hard. In this work, we propose an approach to take the following two constraints into account during the grasp detection stage, namely, (i) the picked object must be able to be placed with a predefined configuration without in-hand manipulation (ii) it must be reachable by the robot under the joint limit and collision-avoidance constraints for both pick and place cases. Our key idea is to train an SE(3) grasp diffusion network to estimate the noise in the form of spatial velocity, and constrain the denoising process by a multi-target differential inverse kinematics with an inequality constraint, so that the states are guaranteed to be reachable and placement can be performed without collision. In addition to an improved success ratio, we experimentally confirmed that our approach is more efficient and consistent in computation time compared to a naive two-stage approach.

cs.RO↗

arXiv2025

ZeroGrasp: Zero-Shot Shape Reconstruction Enabled Robotic Grasping

Robotic grasping is a cornerstone capability of embodied systems. Many methods directly output grasps from partial information without modeling the geometry of the scene, leading to suboptimal motion and even collisions. To address these issues, we introduce ZeroGrasp, a novel framework that simultaneously performs 3D reconstruction and grasp pose prediction in near real-time. A key insight of our method is that occlusion reasoning and modeling the spatial relationships between objects is beneficial for both accurate reconstruction and grasping. We couple our method with a novel large-scale synthetic dataset, which comprises 1M photo-realistic images, high-resolution 3D reconstructions and 11.3B physically-valid grasp pose annotations for 12K objects from the Objaverse-LVIS dataset. We evaluate ZeroGrasp on the GraspNet-1B benchmark as well as through real-world robot experiments. ZeroGrasp achieves state-of-the-art performance and generalizes to novel real-world objects by leveraging synthetic data.

cs.RO↗

arXiv2024

A Planar-Symmetric SO(3) Representation for Learning Grasp Detection

Planar-symmetric hands, such as parallel grippers, are widely adopted in both research and industrial fields. Their symmetry, however, introduces ambiguity and discontinuity in the SO(3) representation, which hinders both the training and inference of neural-network-based grasp detectors. We propose a novel SO(3) representation that can parametrize a pair of planar-symmetric poses with a single parameter set by leveraging the 2D Bingham distribution. We also detail a grasp detector based on our representation, which provides a more consistent rotation output. An intensive evaluation with multiple grippers and objects in both the simulation and the real world quantitatively shows our approach's contribution.

cs.RO↗

arXiv2024

Gravity-aware Grasp Generation with Implicit Grasp Mode Selection for Underactuated Hands

Learning-based grasp detectors typically assume a precision grasp, where each finger only has one contact point, and estimate the grasp probability. In this work, we propose a data generation and learning pipeline that can leverage power grasping, which has more contact points with an enveloping configuration and is robust against both positioning error and force disturbance. To train a grasp detector to prioritize power grasping while still keeping precision grasping as the secondary choice, we propose to train the network against the magnitude of disturbance in the gravity direction a grasp can resist (gravity-rejection score) rather than the binary classification of success. We also provide an efficient data generation pipeline for a dataset with gravity-rejection score annotation. In addition to thorough ablation studies, quantitative evaluation in both simulation and real-robot clarifies the significant improvement in our approach, especially when the objects are heavy.

cs.RO↗

arXiv2024

ViFu: Multiple 360$^\circ$ Objects Reconstruction with Clean Background via Visible Part Fusion

In this paper, we propose a method to segment and recover a static, clean background and multiple 360$^\circ$ objects from observations of scenes at different timestamps. Recent works have used neural radiance fields to model 3D scenes and improved the quality of novel view synthesis, while few studies have focused on modeling the invisible or occluded parts of the training images. These under-reconstruction parts constrain both scene editing and rendering view selection, thereby limiting their utility for synthetic data generation for downstream tasks. Our basic idea is that, by observing the same set of objects in various arrangement, so that parts that are invisible in one scene may become visible in others. By fusing the visible parts from each scene, occlusion-free rendering of both background and foreground objects can be achieved. We decompose the multi-scene fusion task into two main components: (1) objects/background segmentation and alignment, where we leverage point cloud-based methods tailored to our novel problem formulation; (2) radiance fields fusion, where we introduce visibility field to quantify the visible information of radiance fields, and propose visibility-aware rendering for the fusion of series of scenes, ultimately obtaining clean background and 360$^\circ$ object rendering. Comprehensive experiments were conducted on synthetic and real datasets, and the results demonstrate the effectiveness of our method.

cs.CV↗

arXiv2024

DiffusionNOCS: Managing Symmetry and Uncertainty in Sim2Real Multi-Modal Category-level Pose Estimation

This paper addresses the challenging problem of category-level pose estimation. Current state-of-the-art methods for this task face challenges when dealing with symmetric objects and when attempting to generalize to new environments solely through synthetic data training. In this work, we address these challenges by proposing a probabilistic model that relies on diffusion to estimate dense canonical maps crucial for recovering partial object shapes as well as establishing correspondences essential for pose estimation. Furthermore, we introduce critical components to enhance performance by leveraging the strength of the diffusion models with multi-modal input representations. We demonstrate the effectiveness of our method by testing it on a range of real datasets. Despite being trained solely on our generated synthetic data, our approach achieves state-of-the-art performance and unprecedented generalization qualities, outperforming baselines, even those specifically trained on the target domain.

cs.CV↗

arXiv2024

Non-convex optimization problems for maximum hands-off control

The maximum hands-off control is the optimal solution to the L0 optimal control problem. It has the minimum support length among all feasible control inputs. To avoid computational difficulties arising from its combinatorial nature, the convex approximation method that replaces the L0 norm by the L1 norm in the cost function has been employed on standard. However, this approximation method does not necessarily obtain the maximum hands-off control. In response to this limitation, this paper newly introduces a non-convex approximation method and formulates a class of non-convex optimal control problems that are always equivalent to the maximum hands-off control problem. Based on the results, this paper describes the computation method that quotes algorithms designed for the difference of convex functions optimization. Finally, this paper confirms the effectiveness of the non-convex approximation method with a numerical example.

math.OC↗

arXiv2023

GS-Pose: Category-Level Object Pose Estimation via Geometric and Semantic Correspondence

Category-level pose estimation is a challenging task with many potential applications in computer vision and robotics. Recently, deep-learning-based approaches have made great progress, but are typically hindered by the need for large datasets of either pose-labelled real images or carefully tuned photorealistic simulators. This can be avoided by using only geometry inputs such as depth images to reduce the domain-gap but these approaches suffer from a lack of semantic information, which can be vital in the pose estimation problem. To resolve this conflict, we propose to utilize both geometric and semantic features obtained from a pre-trained foundation model.Our approach projects 2D features from this foundation model into 3D for a single object model per category, and then performs matching against this for new single view observations of unseen object instances with a trained matching network. This requires significantly less data to train than prior methods since the semantic features are robust to object texture and appearance. We demonstrate this with a rich evaluation, showing improved performance over prior methods with a fraction of the data required.

cs.CV↗

arXiv2023

A Probabilistic Rotation Representation for Symmetric Shapes With an Efficiently Computable Bingham Loss Function

In recent years, a deep learning framework has been widely used for object pose estimation. While quaternion is a common choice for rotation representation, it cannot represent the ambiguity of the observation. In order to handle the ambiguity, the Bingham distribution is one promising solution. However, it requires complicated calculation when yielding the negative log-likelihood (NLL) loss. An alternative easy-to-implement loss function has been proposed to avoid complex computations but has difficulty expressing symmetric distribution. In this paper, we introduce a fast-computable and easy-to-implement NLL loss function for Bingham distribution. We also create the inference network and show that our loss function can capture the symmetric property of target objects from their point clouds.

cs.CV↗

arXiv2022

Matrix Pontryagin principle approach to controllability metrics maximization under sparsity constraints

Controllability maximization problem under sparsity constraints is a node selection problem that selects inputs that are effective for control in order to minimize the energy to control for desired state. In this paper we discuss the equivalence between the sparsity constrained controllability metrics maximization problems and their convex relaxation. The proof is based on the matrix-valued Pontryagin maximum principle applied to the controllability Lyapunov differential equation.

math.OC↗

arXiv2022

Probabilistic Rotation Representation With an Efficiently Computable Bingham Loss Function and Its Application to Pose Estimation

In recent years, a deep learning framework has been widely used for object pose estimation. While quaternion is a common choice for rotation representation of 6D pose, it cannot represent an uncertainty of the observation. In order to handle the uncertainty, Bingham distribution is one promising solution because this has suitable features, such as a smooth representation over SO(3), in addition to the ambiguity representation. However, it requires the complex computation of the normalizing constants. This is the bottleneck of loss computation in training neural networks based on Bingham representation. As such, we propose a fast-computable and easy-to-implement loss function for Bingham distribution. We also show not only to examine the parametrization of Bingham distribution but also an application based on our loss function.

cs.CV↗

arXiv2022

Sim2Real Instance-Level Style Transfer for 6D Pose Estimation

In recent years, synthetic data has been widely used in the training of 6D pose estimation networks, in part because it automatically provides perfect annotation at low cost. However, there are still non-trivial domain gaps, such as differences in textures/materials, between synthetic and real data. These gaps have a measurable impact on performance. To solve this problem, we introduce a simulation to reality (sim2real) instance-level style transfer for 6D pose estimation network training. Our approach transfers the style of target objects individually, from synthetic to real, without human intervention. This improves the quality of synthetic data for training pose estimation networks. We also propose a complete pipeline from data collection to the training of a pose estimation network and conduct extensive evaluation on a real-world robotic platform. Our evaluation shows significant improvement achieved by our method in both pose estimation performance and the realism of images adapted by the style transfer.

cs.CV↗

arXiv2021

Sparse optimal stochastic control

In this paper, we investigate a sparse optimal control of continuous-time stochastic systems. We adopt the dynamic programming approach and analyze the optimal control via the value function. Due to the non-smoothness of the $L^0$ cost functional, in general, the value function is not differentiable in the domain. Then, we characterize the value function as a viscosity solution to the associated Hamilton-Jacobi-Bellman (HJB) equation. Based on the result, we derive a necessary and sufficient condition for the $L^0$ optimality, which immediately gives the optimal feedback map. Especially for control-affine systems, we consider the relationship with $L^1$ optimal control problem and show an equivalence theorem.

math.OC↗

arXiv2021

Multiple sparsity constrained control node scheduling with application to rebalancing of mobility networks

This paper treats an optimal scheduling problem of control nodes in networked systems. We newly introduce both the L0 and l0 constraints on control inputs to extract a time-varying small number of effective control nodes. As the cost function, we adopt the trace of the controllability Gramian to reduce the required control energy. Since the formulated optimization problem is combinatorial, we introduce a convex relaxation problem for its computational tractability. After a reformulation of the problem into an optimal control problem to which Pontryagin's maximum principle is applicable, we give a sufficient condition under which the relaxed problem gives a solution of the main problem. Finally, the proposed method is applied to a rebalancing problem of a mobility network.

math.OC↗