PyTorchKR Papers

최신 논문

총 16,105편
태그·기간 등 상세 필터

카테고리 탐색

로그인 후 이 카테고리 구독

arXiv:2611.26159v1cs.CV

When Foundation Models Disagree: Representational Heterogeneity for Calibrated Uncertainty in Whole-Slide Image Classification

Pathology foundation models supply strong representations for whole-slide image classification, yet the confidence attached to their predictions remains hard to trust. Prevailing uncertainty estimators act within a single feature space and thereby overlook the uncertainty that the choice of foundation-model representation itself induces. This work introduces CALIFUSE, a calibration-oriented framework that fuses several pathology foundation models for the specific purpose of uncertainty estimation. Departing from accuracy-driven fusion, CALIFUSE reads representational heterogeneity across encoders as an explicit confidence cue. Slide-level decision consensus is coupled with chunk-level evidence agreement quantified through Jensen-Shannon divergence, so that both prediction stability and fine-grained evidence consistency inform the estimate. Foundation-model encoders stay frozen. A task-specific nnMIL classifier is still trained for each FM; CALIFUSE then estimates confidence at inference without training an extra fusion module. Evaluated on seven WSI classification tasks under several foundation-model combinations, the framework sharpens calibration, misclassification detection, and selective prediction relative to both the strongest single-encoder baseline and naive ensemble averaging, all while preserving competitive balanced accuracy.

arXiv:2611.26146v1cs.RO

A Practical Proprioceptive Data Collection Utility for the Unitree GO1

The Unitree Go1 is a widely used platform for legged robotics research, yet collecting proprioceptive data from it is unnecessarily difficult: the Unitree SDK is sparsely documented, and ROS1 (Robot Operating System), the Go1's native middleware, is fully deprecated. We present an open-source graphical utility that enables researchers to reliably record proprioceptive data from the Go1, including joint states, IMU (Inertial Measurement Unit) readings, foot force sensor data, and gyroscope measurements, without requiring ROS expertise. The tool provides a simple GUI (Graphical User Interface) for starting, stopping, and labeling recordings in real time. The contribution of this paper is the collection utility. We characterize the acquisition quality of 51 recordings collected with it, including the achieved sampling rate and repeated samples. As a secondary, deliberately modest check that the data is useful, we train a terrain classifier on it. The results are weak: a Temporal Convolutional Network reaches about 38% accuracy on held-out recordings across ten terrain classes (10% would be chance), and a window-level split that ignores recording boundaries inflates the same measurement to about 99%. We report this as a demonstration that the logged signals carry terrain information, not as a competitive classifier. The utility is designed to lower the barrier to collecting Go1 proprioceptive data without ROS.

arXiv:2611.26145v1cs.CV

Can World Models Predict Rice Growth? Toward Agricultural World Models

World models can imagine how a physical scene evolves over the next few seconds, but growers need to know how a crop will develop over weeks to months: will this rice plant have headed a month from now, and how will it look at harvest? We introduce agricultural world modeling, the task of predicting how the same plant will look at a requested horizon of days to months, from one photo of it, its cultivar, and its camera view. In preliminary analysis, even state-of-the-art image generators get this wrong: they often grow the plant too little or give it the color and form of the wrong growth stage. We argue that they do not work out the plant's developmental state, which the photo reveals only indirectly. We therefore propose AgroWM, which reasons about development before it draws growth. Its reasoning module, PhenoClock, works out the current growth stage, predicts the future one, and writes a Growth Script with numeric targets for height, plant area, and color; its generation module, the Rooted Renderer, then edits the photo to follow this script, adding new organs and colors to the photographed plant. Because pixel scores can prefer an unchanged copy of the photo, we instead check each forecast's traits and developmental timing against the plant's own later photos. On rice, 54.6% of the forecasts of AgroWM are matched to the target stage (Stage R@1), against 34.8% for GPT Image 2.5, 24.0% for Qwen Image 2.1, and 47.4% for the strongest domain-specific baseline, the same backbone fine-tuned without Growth Scripts; 9 to 20 weeks ahead, it still reaches 46.4%, against 35.7% for that baseline. On maize, a crop with a different body plan and stage scale, the same design again scores best on all four measures.

arXiv:2611.26128v1cs.CV

CryoGaussian: Heterogeneous Cryo-EM Reconstruction Made Explicit with Gaussian Splatting

Single-particle cryo-electron microscopy (cryo-EM) determines macromolecular structures from noisy particle images. However, rare, transient compositional states and subtle, continuous conformational variation are difficult to resolve. We present CryoGaussian, a Gaussian splatting method for heterogeneous cryo-EM reconstruction that fits an explicit Gaussian density representation directly to the particle images through a differentiable rasterizer rendering orthographic line integrals modulated by the contrast transfer function. The Gaussians are initialized randomly, with their number derived from the estimated spatial support of the molecular density and their minimum size bounded by the Nyquist frequency of the particle images. To capture the compositional and conformational variation across particle images, we learn particle-image and per-Gaussian latents that jointly modulate the base position, amplitude, and scale of each Gaussian. Each structural state is then reconstructed by fitting the Gaussians conditioned on its inferred latent. On CryoBench, CryoGaussian achieves state-of-the-art reconstruction accuracy for both compositional and conformational heterogeneity. On experimental data, it resolves a rare compositional state of the 50S ribosomal subunit (class C4, 1.3% of particles) with rRNA helix 68 docked in the absence of the central protuberance, and continuous conformational motion of the SF3b and helicase subcomplexes of the pre-catalytic spliceosome. We further optimize the rasterizer for additive cryo-EM projection, reducing training time by 12.8-28.0$\times$.

arXiv:2611.26120v1cs.CV

Better Retrieval, Not Better Binding: Pretraining Exposure in Egocentric Video Retrieval

Video-language models (VLMs) achieve strong performance on video-text retrieval and are increasingly used for egocentric video understanding. Yet, despite gains in zero-shot retrieval, they have shown limitations in compositional understanding, and it remains unclear whether they understand compositional relationships between verbs and nouns. In this work, we introduce Ego-ExBind, a compositional diagnostic protocol for analyzing pretraining exposure and verb-noun binding in egocentric video-text retrieval. Ego-ExBind consists of an Exposure Ledger built from the EgoClip pretraining corpus used by EgoVLPv2, which defines Seen Composition, Unseen Composition, and Unseen Atom splits for EPIC-KITCHENS-100 multi-instance retrieval (MIR). We further probe noun and verb binding channels through controlled noun-swap and verb-swap negatives. We find that zero-shot retrieval is positively associated with pair exposure within seen compositions, even after controlling for the marginal frequencies of verbs and nouns. However, higher pair exposure is associated with larger noun-swap margins and slightly smaller verb-swap margins. This asymmetry motivates us to test whether exposure reweighting can improve retrieval with little change in the measured swap margins. To isolate the effect of exposure input, we keep the trained checkpoint fixed and replace only the exposure input at inference. We demonstrate that this exposure reweighting can improve video-text retrieval without stronger verb-noun compositional understanding, and that the gain depends on the correct sample-level exposure correspondence. By separating exposure-driven retrieval gains from verb-noun binding, Ego-ExBind provides a diagnostic framework for evaluating whether future improvements in egocentric MIR reflect compositional understanding rather than pretraining familiarity.

arXiv:2611.26116v1cs.CV

Learning Reconstruction Errors as Spatiotemporal Evidence for AI-Generated Video Detection

The proliferation of AI-generated videos has raised growing concerns about misinformation, fraud, and potential societal risks, creating an urgent need for detection. Reconstruction-error-based methods have shown strong empirical performance for AI-generated image detection by exploiting discrepancies between an input and its reconstructed version. Their extension to video detection, however, faces two challenges: (i) a fixed reconstructor is not optimized for the detection objective and may yield weak discriminative signals; and (ii) an additional raw-video feature stream is often introduced to model temporal dependencies, increasing architectural overhead. To address these challenges, we develop a detector that learns and exploits reconstruction errors as a unified source of spatiotemporal evidence. First, we jointly optimize a reconstructor and the downstream classifier to produce reconstruction errors that are more informative for detecting AI-generated videos. Second, we show that consecutive reconstruction errors alone carry temporal evidence, and thus a 3D convolution over them captures spatiotemporal evidence without a raw-video stream. Extensive experiments on over 30 existing AI-video generators demonstrate competitive performance against baselines.

arXiv:2611.26107v1cs.CV

HiUni: Harmonizing Understanding and Generation in Shared Visual Space

Unified multimodal models aim to support visual understanding and generation within a single framework. However, prevailing approaches often separate their visual representations or backbone computation to accommodate the different requirements of these two tasks, which limits the degree of unification. In this paper, we present HiUni, which revisits this design choice and explores whether a pretrained vision language model can be directly extended into a unified model with extensive parameter sharing. Specifically, HiUni reuses the same language-aligned ViT and dense Transformer backbone for both capabilities, while confining task-specific computation to the output side. Yet such a design introduces two key challenges. Generation in pixel space is difficult to optimize, and semantic abstraction in pretrained ViT discards fine spatial details required for synthesis. To address these challenges, HiUni introduces a \emph{Pixel Diffusion Head} tailored to direct pixel-space optimization and a lightweight \emph{HiGate} pathway that provides fine-grained spatial details near the output stage. Extensive experiments show that HiUni achieves a similar average understanding score to the pretrained VLM while attaining competitive performance on image generation and editing. Controlled comparisons and analyses further demonstrate that a single dense backbone can effectively support both capabilities without task-specific backbone computation, highlighting a simpler path toward unified multimodal modeling.

arXiv:2611.26103v1cs.RO

UniVEP: Unifying Velocity and Endpoint Prediction for Robot Manipulation

Flow-matching robot policies usually use one prediction target throughout action generation, even though recovering actions from noise and polishing nearly clean actions bring different challenges. We systematically compare velocity prediction (V), endpoint prediction (X), and velocity-weighted endpoint prediction (X--V) across multiple robot policies, benchmarks, and inference budgets. On StarVLA and LiLa-WAM, X--V works best for few-step generation, while V gains more from iterative refinement. We further uncover their complementary strengths across noise levels: clean actions have concentrated covariance spectra, endpoint predictions stay more stable under high noise, and velocity predictions offer more accurate corrections near clean actions. Inspired by these findings, we introduce UniVEP, which learns both prediction targets together with consistency supervision and switches from endpoint recovery to velocity refinement as noise drops. Each step only needs one network run, and joint training also boosts single-step endpoint generation. Tests on simulated and real-world manipulation show better success rates and efficiency. On RoboTwin, UniVEP hits 61.36% success with one StarVLA run, beating the best single-target baseline at sixteen runs (59.96%). With LiLa-WAM, two runs outperform the best sixteen-step single-target baseline and cut inference latency by 77.1%.

arXiv:2611.26098v1cs.CV

Frequency-Synchronized Boundary Coupling for Training-Free Multi-Prompt Long Video Generation

Text-to-video generation has made remarkable progress with diffusion models and transformer-based video architectures, yet generating coherent long videos from evolving textual descriptions remains challenging. Training long-video models from scratch requires substantial computation and large-scale multi-event data, while training-free multi-prompt generation must activate new prompt semantics without breaking temporal continuity. We identify a frequency-agnostic boundary coupling problem in existing transition strategies: weak coupling causes abrupt changes, flicker, and background jumps, whereas overly strong or coarse coupling may over-preserve source semantics and produce ghosting or duplicated subjects around prompt boundaries. To address this problem, we propose FreqSync, a training-free framework that formulates prompt transitions as frequency-selective boundary coupling. FreqSync combines Frequency-Synchronized Source-Conditioned Attention (FS-SCA) with Residual-Clipped Spectral Stitching (RCSS), enabling adjacent segments to share stable low- and mid-frequency structure and motion cues while keeping high-frequency prompt-specific details target-dominant under explicit residual constraints. Extensive experiments show that FreqSync achieves state-of-the-art transition smoothness and prompt alignment while maintaining strong perceptual video quality, demonstrating its effectiveness for coherent multi-prompt long video generation.

arXiv:2611.26072v1cs.CV

DfD-GS: Defocus Image Prior for Metric-Scale Single-View Depth Estimation and View Synthesis

Recovering metric-scale geometry from observations at a single viewpoint remains fundamentally ill-posed, limiting the reliability of novel view synthesis. Recent methods rely on learned monocular depth priors, which provide strong structural estimates but can suffer from scale ambiguity and dataset-dependent biases. We propose DfD-GS, a feed-forward framework that predicts a metric-scale 3D Gaussian scene representation from a single-view all-in-focus--defocused image pair. The key idea is to exploit defocus blur as a physical metric depth cue. DfD-GS first estimates a defocus-based metric disparity map from the all-in-focus--defocused image pair and the corresponding defocus strength. This estimate is used to align and refine a monocular depth prior to metric scale. Visible and hidden 3D Gaussians are then predicted for occlusion-aware novel view synthesis. Experiments on synthetic-defocus datasets constructed from Booster and Middlebury show that DfD-GS outperforms single-image baselines in metric depth estimation and novel view synthesis. On WildRGB-D, DfD-GS achieves lower AbsRel and LPIPS but higher depth RMSE than the single-image baselines. Additional analyses show that DfD-GS maintains a more stable metric scale under field-of-view changes. Qualitative results on a real dataset show recovery of the main scene structure under optical defocus.

arXiv:2611.26065v1cs.CV

Domain-Aware Conformal Prediction for Trustworthy Spacecraft Pose Estimation under Sim-to-Real Shift

Autonomous rendezvous and proximity operations require not only an accurate estimate of the pose of the target spacecraft, but also a reliable indication of when this estimate can be trusted. Conformal prediction provides such an indication with a formal guarantee, by producing pose uncertainty sets that contain the true pose with a specified probability. Such a guarantee, however, assumes that calibration and test data follow the same distribution, an assumption that is violated in orbital perception, where labelled calibration images are available only from synthetic renderings while operational images are real. In this work, we show on the SPEED+ dataset that conformal pose estimation calibrated on synthetic images loses its coverage guarantee on real images, and that many of the resulting failures are silent: the predicted pose set appears valid but excludes the true pose. To address this, we calibrate on synthetic images translated to the appearance of real images. Since translation alters appearance while preserving geometry, the synthetic pose labels remain valid and no labelled real images are required. Experiments with two translation models show that this calibration restores coverage above the specified level for miscoverage levels of 0.1 and above, and reduces silent failures from up to 43.5% to at most 3.3% of the test images, bringing conformal pose estimation a step closer to trustworthy deployment in space.

arXiv:2611.26058v1cs.CV

One-to-Many: Diverse and Bidirectional Image-to-Image Translation via Rectified Flow

Many paired image-to-image translation tasks admit multiple valid target appearances, while training provides only one target per source. We propose One-to-Many Flow (O2MFlow), a latent-conditioned Rectified Flow (RF) model for one-to-many generation and backward translation (target$\rightarrow$source) within a single transport framework. Conditioning the RF dynamics on an appearance latent turns transport into a family of latent-indexed bijections: a fixed latent gives a mapping that is inverted by time-reversed integration, and sampled latents produce diverse outputs for the same input. We formalize reusable appearance variations as Shared Latent Transport and show that Kullback--Leibler (KL) regularization directly controls the source--latent distribution gap between training and generation. Across three benchmarks, O2MFlow offers dataset-dependent forward diversity--fidelity trade-offs and the lowest backward LPIPS-Dist among evaluated methods, with the lowest backward FID on two benchmarks.

arXiv:2611.26056v1cs.CV

Less Than Meets the Eye: Action-Set Identifiability and Principled Abstention for 3D Physical Geometry

A corridor, its mirror reflection, and its image on a display can look identical, yet their physical surfaces lie at different depths. We test whether bounded camera motions distinguish these explanations. Intervene3D takes an image and an initial geometry estimate, predicts the view under each hypothesis and candidate motion, and selects a motion by maximin separability. It reports the mechanism and first physical surface only if the available motions separate the hypotheses still in contention. Otherwise it abstains. An adequacy test also rejects observations that no hypothesis explains. On a matched synthetic benchmark with exact landmarks, disabling abstention gives confident answers on 76.4% of unresolvable scenes. With abstention, none receives a confident answer under exact or noisy landmarks. At 100% accuracy on answered scenes, maximin selection answers 61.0% of scenes and summed selection 43.1%. The adequacy test lowers wrong answers on strongly curved mirror scenes from 56.8% to 12.4%. Real display experiments compare two hypotheses using one frozen depth-network estimate. The second image supplies photometric evidence and is never passed through the depth network. One executed motion reduces the pooled median surface error of twelve networks by 31.5%. These networks have larger error inside the annotated panel than outside on 33.4% of illusion-benchmark images. They place the first physical surface behind the glass on 36.9% of Mirror3D frames.

arXiv:2611.26045v1cs.CV

PaveSplat: Resolution-Robust Feed-Forward Gaussian Splatting with Residually Corrected Surface-Aligned Covariances

We present PaveSplat, a feed-forward Gaussian splatting method that adds a learned residual to the analytic surface-aligned covariance, keeping surfels paving the visible surface across rendering resolutions. Pixel-aligned feed-forward Gaussian splatting predicts one primitive per pixel from a few images in a single pass, and its rendering quality depends on how well these primitives cover the surface. Recent surfel-based methods compute the surface-aligned covariance analytically from predicted depth. However, this covariance has no way to correct the surfel orientation where the depth-derived frame is ill-posed. The surfels then cover the surface loosely; at the training resolution each surfel's one-pixel minimum footprint covers the gaps, but they show up as holes when the same prediction is re-rendered at a higher resolution. To address this, we keep the analytic covariance as an initialization, with its axes aligned to the surface tangents, and add an identity-initialized unit-quaternion residual that reorients a surfel only where this improves rendering, turning surfels that the analytic frame mis-aims back onto the surface. On RealEstate10K, PaveSplat achieves the best PSNR, SSIM, and LPIPS among single-Gaussian-per-pixel methods at every scale (1x-8x). The native-resolution gain is small; the benefit appears when the same prediction is rendered at higher display resolutions without re-running the network. On zero-shot DL3DV, it leads PSNR within this group at every scale and ranks first or second on SSIM/LPIPS. A coverage-hole analysis shows fewer uncovered pixels than both baselines under high-resolution re-rendering.

arXiv:2611.26040v1cs.CV

Streaming-AAA: Streaming AAA Game Video Generation at High Resolution, High Frame Rate, and Low Latency

Controllable streaming video generation offers a promising route toward interactive game simulation, yet existing methods remain far from practical gameplay. They often fail to jointly achieve high visual fidelity, high frame rates, and low interaction latency, resulting in poor responsiveness and degraded user experience. We present Streaming-AAA, a new paradigm for controllable infinite generation of AAA-quality game videos. Our system generates 1080p gameplay videos at 43 FPS with 198 ms end-to-end latency interaction on merely NVIDIA A800 GPUs, while supporting long-horizon play without severe temporal degradation or collapse. Streaming-AAA is built upon three key designs. First, we redesign the VAE architecture for efficient high-resolution decoding, enabling real-time synthesis at 1080p. Second, we pretrain the diffusion model using diffusion forcing and significantly extend the temporal training horizon, allowing the model to capture long-term dependencies and maintain persistent memory. Third, we perform post-training distillation and memory compression to improve inference efficiency while retaining visual quality and temporal consistency. Experiments show that Streaming-AAA substantially improves visual fidelity, frame rate, latency, and long-term stability over prior controllable video generation approaches, enabling a significantly more playable experience for interactive AAA game video generation.

arXiv:2611.26035v1cs.CV

Fixing the Twin, Not the Detector: Sim-to-Real Roadside LiDAR Detection from Digital-Twin Data

Roadside LiDAR detectors are expensive to train on real labels, and digital twins promise to replace these labels, but a detector trained on a twin meets a different sensor, different objects and different annotation conventions on the real site. We describe the pipeline that placed first in the LUMPI track of the UrbanTwin Sim2Real LiDAR Challenge (IEEE BigData 2026 Cup, 0.5541 combined score) and fifth on its V2X-Real track (0.8102). Every detector is a CenterPoint model trained only on synthetic data, on a single 8 GB laptop GPU. Most of the gain came from correcting differences between the twin and the real site rather than from the detector. We align the twin's coordinates, point density and box convention with the test sensor, generate the pedestrians and riders that the twins lack, plus street clutter as unlabeled negatives, and rebuild twin objects at real size and real point density. A non-learned resampling to the per-cell point density of unlabeled real frames lifts LUMPI realism from 0.589 to 0.947, against 0.949 between real frames. Label-free post-processing then exploits what a fixed roadside sensor guarantees: static structures recur in every frame, and real objects have characteristic point statistics. Reporting the validation and test gain of each component side by side, we find that corrections of systematic sim-to-real differences transfer to the test set, typically keeping a quarter to three quarters of their validation gain, while choices that fit the few objects of a validation set do not.

arXiv:2611.25980v1cs.RO

Learning What Actions Change: Action Effect Matching for Dynamics-Faithful World Models

World models aim to predict environment dynamics under actions, enabling the prediction of action consequences without executing them in the real world. Yet despite producing increasingly realistic videos, current world models often fail to capture how small action changes affect future observations. This limits their use in tasks such as robotic policy evaluation and model-based planning, where nearly identical actions can lead to qualitatively different outcomes. To address this, we propose AFFECT, an auxiliary training objective for learning the distribution of visual changes induced by action perturbations. Specifically, we inject temporally correlated noise into the actions of a policy trained on task-agnostic human teleoperation. We then train the world model so that its generated videos respond to this noise as the recorded videos do, without requiring repeated executions from the same initial state. Across OGBench simulation and real-world DROID and ALOHA-2 setups, AFFECT makes world-model rollouts more faithful to real outcomes. On DROID, the Pearson correlation between policy success rates in the world model and on the real robot rises from 0.47 to 0.79, alongside up to 23% lower FVD. These gains carry over to planning, where selecting actions with AFFECT-trained world models improves task success by up to 8.3%.

arXiv:2611.25979v1cs.CV

Multi-View Transformer with Geometry-Guided Multi-Modal ControlNet for 3D Virtual Try-On

Editing a 3D Gaussian Splatting (3DGS) capture of a person so that they wear a new garment requires every rendered view to agree on the same garment details. Existing 3D virtual try-on (VTON) pipelines, however, edit training views largely independently and fuse them with hand-designed heuristics such as manually chosen reference views and fixed attention-blending weights. The resulting inconsistencies are baked into the 3D representation and surface as texture drift, blurred prints, and implausible geometry in novel views. We present GS-VTON-MVT, a three-stage framework that combines implicit and explicit 3D reasoning for consistent 3DGS try-on. (i) A multi-view inpainting transformer jointly edits a set of training views with full cross-view self-attention, so that cross-view correspondence is learned rather than imposed. (ii) A geometry-guided multi-modal ControlNet conditions the diffusion editor on the garment image together with depth and surface normals rendered from the current 3DGS state, supplying explicit structure where visual evidence is weakest. (iii) A pose-adaptive persona-aware attention replaces the fixed blending weight between current-view and reference-view features with weights derived from camera-pose proximity. On the 3D-VTONBench benchmark, GS-VTON-MVT improves FID, KID, LPIPS-based cross-view consistency, and user preference over general-purpose 3D editors and prior 3D try-on methods.

arXiv:2611.25975v1cs.CV

Sample-efficient end-to-end Optical Music Recognition through Active Learning

End-to-end approaches to Optical Music Recognition (OMR) have recently shown promising results, but their performance often degrades on unseen data. A common solution is to pretrain and fine-tune models with annotated samples from the target domain, improving adaptation to new distributions. However, this process typically requires extensive manual annotation to yield competitive models, making it costly and time-consuming. In this work, we explore Active Learning to reduce this burden by selecting informative samples for labeling in low-resource scenarios, where the amount of manual annotation that can be performed is limited. We evaluate both uncertainty- and diversity-based sampling strategies on staff-level and full-page corpora. Results show that confidence-based sampling performs comparably to oracle selection (based on true error), and that each scenario benefits from a tailored strategy.

arXiv:2611.25963v1cs.CV

MorphTwin: Enriching Deformable Material Responses from Videos for Morphable and Interactive Physical Twins

Reconstructing interactive physical twins from video requires accurately capturing how objects deform while maintaining efficient simulation. We observe that reproducing motion in a video does not guarantee correct responses to new forces when the physical representation restricts how an object can deform. In spring-based twins, coupling between volume and shape responses creates such a limitation. We present MorphTwin, which enriches local deformation modes to capture and independently control these responses while retaining standard spring forces and explicit simulation. To support interpretable material editing and transfer, we use offline calibration to map shared relative response changes to stiffness updates on each reconstructed graph. Under identical inverse settings on synthetic scenes, MorphTwin reduces the mean absolute error of Poisson's ratio by 75.8% and the mean relative error of Young's modulus by 22.1% compared to PhysTwin. We introduce an egocentric stereo and RGB-D benchmark covering interactions with objects of varying stiffness. Real-world experiments demonstrate improved reconstruction accuracy and up to 14.4% lower future tracking error than PhysTwin at interactive speeds.