arXiv subjects

Chin-Tien Wu

Publications and source records attributed to Chin-Tien Wu.

6 recordsLinked to original sources

arXiv2026

AstraMoE-SR: Trajectory-Guided Diffusion for Blind Satellite Jitter Deblurring and Super-Resolution

Pushbroom satellite imaging couples limited spatial resolution with platform attitude instability. Platform jitter produces spatially varying motion blur because each scan line is acquired under a different instantaneous attitude, while perspective geometry causes the same perturbation to induce different pixel displacements across the field of view. Existing blind restoration methods that assume a spatially invariant kernel and satellite jitter correction methods that rely on auxiliary observations are therefore not directly applicable. We present AstraMoE-SR, a single-image framework that jointly restores motion blur and spatial resolution without auxiliary measurements. Rather than estimating a blur kernel, we infer how the camera moved by reparameterizing degradation as a local exposure trajectory under pushbroom geometry. A conditional diffusion model estimates the trajectory distribution, mitigating the over-smoothing of high-frequency jitter by deterministic point estimation. The predicted trajectory conditions a pretrained latent diffusion backbone through trajectory-guided geometric alignment and spatially adaptive reconstruction. We further show that the remaining point-wise trajectory error is consistent with intrinsic jitter-phase ambiguity that is not resolved by increasing estimator capacity. On all 1,411 DOTA-v1.0 images degraded using our physically motivated forward model, AstraMoE-SR is the only evaluated method to outperform the no-restoration baseline across every fidelity metric, improving on StableSR by 0.64 dB PSNR, 15.2% LPIPS, and 0.091 DINO feature similarity. Reconstructions conditioned on predicted trajectories differ negligibly from those using ground-truth trajectories, indicating that the estimates retain the degradation information required for effective restoration.

eess.IV↗

arXiv2026

ReynoldsFlow: Physics-Inspired Spatiotemporal Flow Representation for Video Understanding

Video understanding has largely relied on deep spatiotemporal architectures, including 3D convolutional networks and optical flow (OF) based models. While effective, these methods are often computationally expensive and depend on heuristic motion representations that are sensitive to illumination, scale, and structural changes. To address these limitations, we propose ReynoldsFlow, a physics-inspired representation grounded in the Reynolds transport theorem (RTT) and Helmholtz-Hodge decomposition (HHD). ReynoldsFlow decomposes motion into curl-free (CF) and divergence-free (DF) components, providing a principled and interpretable characterization of scene dynamics. By coupling intensity information with decomposed motion cues, it produces dynamics-aware, texture-preserving features that boost downstream tasks such as pose estimation, action recognition, and tiny object detection. Lightweight and modular, ReynoldsFlow can be readily integrated into existing architectures. Experiments across diverse benchmarks show that ReynoldsFlow consistently matches or surpasses existing approaches, offering improved generalizability and computational efficiency.

cs.CV↗

arXiv2026

A Three-Stage Offline SDRE-Based Control Framework for Human Motion Reproduction on a Suspended Bipedal Robot

Evaluating lower limb exoskeletons directly with human subjects can expose users to risk when actuator faults, joint misalignment, or unsuitable assistance occur. Therefore, captured human motion must first be converted into commands that are executable by the robot hardware and repeatable across trials. This paper presents a three-stage offline command generation framework for reproducing lower limb motion and torque on a suspended bipedal robot platform used as a robotic bench system for exoskeleton evaluation. First, State-Dependent Riccati Equation control is applied to the robot dynamic model to obtain a reference torque trajectory associated with measured lower limb motion. Second, parameterized optimization converts this reference into trapezoidal joint velocity commands subject to motor speed and acceleration limits. Third, a proportional-integral-derivative linear quadratic regulator (PID-LQR) compensation refines the command profiles using experimental tracking data. Walking and squatting motions recorded by a Vicon motion capture system are reproduced on the suspended robot to evaluate tracking accuracy and repeatability. The results show that the average root mean square error (RMSE) and standard deviation (STD) of joint angles across repeated trials remain below 3° and 0.15°, respectively. Comparisons of joint angles and torques further show that the proposed method achieves lower maximum RMSE and STD values than the two baseline controllers in all reported cases. These results indicate that the proposed three-stage control provides repeatable and actuator-feasible motion reproduction on a suspended bipedal robot platform as a preliminary test environment for lower limb exoskeleton research before tests involving human subjects.

cs.RO↗

arXiv2023

Spectral Analysis for Semantic Segmentation with Applications on Feature Truncation and Weak Annotation

It is well known that semantic segmentation neural networks (SSNNs) produce dense segmentation maps to resolve the objects' boundaries while restrict the prediction on down-sampled grids to alleviate the computational cost. A striking balance between the accuracy and the training cost of the SSNNs such as U-Net exists. We propose a spectral analysis to investigate the correlations among the resolution of the down sampled grid, the loss function and the accuracy of the SSNNs. By analyzing the network back-propagation process in frequency domain, we discover that the traditional loss function, cross-entropy, and the key features of CNN are mainly affected by the low-frequency components of segmentation labels. Our discoveries can be applied to SSNNs in several ways including (i) determining an efficient low resolution grid for resolving the segmentation maps (ii) pruning the networks by truncating the high frequency decoder features for saving computation costs, and (iii) using block-wise weak annotation for saving the labeling time. Experimental results shown in this paper agree with our spectral analysis for the networks such as DeepLab V3+ and Deep Aggregation Net (DAN).

cs.CV↗

arXiv2017

A Novel Stretch Energy Minimization Algorithm for Equiareal Parameterizations

Surface parameterizations have been widely applied to computer graphics and digital geometry processing. In this paper, we propose a novel stretch energy minimization (SEM) algorithm for the computation of equiareal parameterizations of simply connected open surfaces with a very small area distortion and a highly improved computational efficiency. In addition, the existence of nontrivial limit points of the SEM algorithm is guaranteed under some mild assumptions of the mesh quality. Numerical experiments indicate that the efficiency, accuracy, and robustness of the proposed SEM algorithm outperform other state-of-the-art algorithms. Applications of the SEM on surface remeshing and surface registration for simply connected open surfaces are demonstrated thereafter. Thanks to the SEM algorithm, the computations for these applications can be carried out efficiently and robustly.

cs.GR↗

arXiv2015

Conformal Surface Morphing with Applications on Facial Expressions

Morphing is the process of changing one figure into another. Some numerical methods of 3D surface morphing by deformable modeling and conformal mapping are shown in this study. It is well known that there exists a unique Riemann conformal mapping from a simply connected surface into a unit disk by the Riemann mapping theorem. The dilation and relative orientations of the 3D surfaces can be linked through the Möbius transformation due to the conformal characteristic of the Riemann mapping. On the other hand, a 3D surface deformable model can be built via various approaches such as mutual parameterization from direct interpolation or surface matching using landmarks. In this paper, we take the advantage of the unique representation of 3D surfaces by the mean curvatures and the conformal factors associated with the Riemann mapping. By registering the landmarks on the conformal parametric domains, the correspondence of the mean curvatures and the conformal factors for each surfaces can be obtained. As a result, we can construct the 3D deformation field from the surface reconstruction algorithm proposed by Gu and Yau. Furthermore, by composition of the Möbius transformation and the 3D deformation field, the morphing sequence can be generated from the mean curvatures and the conformal factors on a unified mesh structure by using the cubic spline homotopy. Several numerical experiments of the face morphing are presented to demonstrate the robustness of our approach.

cs.GR↗