ML Reads

Personal arXiv list

ML papers to read today.

Pick a topic and keep a small daily list of papers worth opening.

Refresh queueDaily mix

Today's queue

5 papers

#01Nov 16, 2026

cs.CV

MetricHand: Geometry-Aware Flow Matching for 4D Hand Reconstruction in Metric World Space

Yibo Bai, Likui Zhang, Jinglei Zhang and 4 more

Egocentric hand motion provides physical motion supervision for dexterous manipulation and vision-language-action pretraining. Reliable use of this supervision requires accurate and temporally coherent 4D hand reconstruction in metric world space. Existing methods typically lift framewise estimates produced under weak perspective into world coordinates. By ignoring joint-dependent depth variation at close range, this approximation couples errors in translation, orientation, and articulation that persist after world lifting and worsen under occlusion. We present MetricHand, a geometry-aware flow matching framework for 4D hand reconstruction in metric world space. MetricHand encodes depth priors from hand-object interaction regions into joint-associated representations that preserve local spatial structure and metric scale. Fused with visual features, this geometry guides coarse metric estimation and conditions every step of temporal generation. The model jointly recovers world-space translation, global orientation, MANO articulation, and shape, while structural priors and temporal context support estimation under partial occlusion. MetricHand reduces camera-space MPJPE by 49.3% and 52.9%, absolute world-space MPJPE by 28.7% and 21.7%, and wrist localization error by 48.5% and 37.9% relative to HaWoR, respectively. World-space acceleration error decreases by 44.2% and 64.3%, demonstrating improved temporal fidelity alongside more accurate absolute localization.

#02Nov 16, 2026

cs.CV

Oriented Semi-DETR for Semi-Supervised Aerial Object Detection

Yuqiu Li, Jian Xue, Fan Li and 1 more

Semi-supervised object detection has recently been extended to DETR, where oriented boxes introduce an ambiguity that does not arise for axis-aligned boxes: the map from parameters to rectangles is two-to-one, since $(c_x,c_y,w,h,θ)$ and $(c_x,c_y,h,w,θ+π/2)$ denote the same rotated rectangle. Losses and matching costs are computed from the parameters, so two boxes that coincide in space can be far apart in the metric that drives training. A correctly localized student box is then penalized by the regression loss, and the inflated matching cost leads the Gaussian mixture model (GMM) filter to discard it. We propose Oriented Semi-DETR, which removes this ambiguity wherever a pseudo box enters training. Swap-aligned Target Selection regresses towards the encoding closer to the student prediction. A Swap-invariant Reliability Cost minimizes over the two encodings inside the matching cost, so the filter decides on geometry alone. Swap-aware Pseudo-label Denoising starts denoising earlier, ramps its loss weight linearly, and weights each query by the teacher score and encoding margin. All three are parameter-free and require no additional forward pass. Across all label ratios of DOTA-v2.0, Oriented Semi-DETR outperforms DETR-based supervised and semi-supervised baselines, with the largest gains at the strict IoU threshold.

#03Nov 16, 2026

cs.CV

AnachroBench: Evaluating Historical Consistency in Large Vision-Language Models

Keunhyeung Park, Byeonggeuk Lim, Kahyeon Kim and 5 more

Existing Large Vision-Language Model benchmarks primarily evaluate visual understanding in general everyday scenarios, leaving historical consistency with a specified era and country largely underexplored. To address this gap, we propose AnachroBench, a controlled discriminative benchmark for evaluating historical consistency in visual scenes. It contains 1,600 image pairs from eight countries spanning the late 19th to the early 20th century, each pairing an original historical image with an anachronistic version containing a seamlessly inserted modern object. AnachroBench evaluates three aspects: Inserted Object Perception, which tests whether models detect the inserted modern object; Era-Country Consistency, which assesses whether the scene is consistent with the specified era and country; and Conditional Era-Country Consistency, which measures consistency judgments only when the inserted object is correctly perceived. A false consistency judgment occurs when a model perceives the modern object but still judges the scene as historically consistent. Overall, AnachroBench reveals a gap between inserted object perception and era-country consistency judgments, providing a targeted evaluation of this behavior.

#04Nov 16, 2026

cs.CV

IGRecon4D: Dynamic-Object-Aware Feed-Forward Reconstruction of Driving Scenes

Yifei Tong, Ziyang Huang, Yizhao Han and 2 more

Feed-forward 4D driving-scene reconstruction enables efficient modeling of dynamic autonomous-driving scenes, but current methods suffer from two complementary failure modes: semantic leakage of traffic participants into the static representation and geometric drift of Gaussian locations across viewpoints. We propose \textbf{IGRecon4D}, a feed-forward dynamic scene reconstruction framework that addresses these failures with training-time instance-prior distillation and LiDAR-anchored Gaussian positioning. Specifically, an instance-level semantic branch provides a potentially dynamic participant prior, which is distilled into dynamic-mask learning through the Instance-Dynamic Consistency Constraint to encourage high-recall localization of traffic-participant regions prone to dynamic-static leakage. In parallel, sparse LiDAR measurements guide dense depth estimation, and the resulting depths are back-projected to anchor Gaussian positions for improved multi-view geometric consistency. Experiments on the Waymo Open Dataset and zero-shot evaluation on nuScenes demonstrate improved full-image reconstruction on the main benchmark and stronger traffic-participant-region reconstruction under cross-dataset evaluation.

#05Nov 16, 2026

cs.CV

GRAZE: Text-Free Grasp Synthesis for Unseen Objects from Human Gaze

Jeongho Lee, Boeun Kim, Hyung Jin Chang and 2 more

Text-conditioned hand-object interaction (HOI) generation provides intuitive semantic control, but language is an indirect way to specify where an interaction should occur on an object. Naming the part to grasp requires object-specific words, and naming the approach side requires a reference frame that a prompt lacks. We propose GRAZE (GRAsp from gaZE), which generates hand-object grasp motion from recorded gaze instead of object and part names. Its Geometry-Aware Gaze Encoding (GeoGaze) maps gaze rays to cues on object geometry, and cross-attention conditions a diffusion Transformer on these cues. We also introduce HOT3D-Grasp, a grasp-centered benchmark with object-disjoint folds built on HOT3D recordings. On unseen objects, GRAZE grasps the target part more often than the strongest text-conditioned baseline (49.3% vs. 43.6%), achieves the highest grasp success (66.3% vs. 46.6%), and transfers without fine-tuning to GazeHOI (45.3% vs. 33.8%). Gaze conveys the approach side: when gaze is rotated around the object, the approach turns with it in 59.1% of cases, against 20.5% for a text model retrained with clock words. On the same backbone, a text prompt names parts better, but gaze adds the approach side. Given both, GRAZE keeps the prompt's part accuracy and follows gaze.