arXiv：cs.AI· Hongyu Ma, Hairong Qu, Shiqi Zhao, Yongsong Yang, Peng Yin·· 6 小时前
ESTHER：面向可穿戴设备的自我中心立体手部估计与重建模型及 ESTHER3D 基准
ESTHER: Egocentric Stereo Hand Estimation and Reconstruction in the Wild
AI 导读
ESTHER 是面向可穿戴自我中心立体视觉的手部重建模型，其立体几何、时序推理与输出表示均针对该信号设计，通过校准标注流程生成的伪标签训练。它同时构建了 ESTHER3D 基准，将野外训练集与动捕真实度量测试集配对，实验取得 state-of-the-art 精度、更强的外部泛化能力，并能应对缺失视图、丢帧和光照运动模糊等真实采集问题。
正文
Abstract:Human dexterity is guided by two eyes watching two hands: binocular vision supplies the metric 3D structure that fine-grained manipulation consumes. Egocentric stereo is therefore the natural perceptual interface for robots, AR, and VR-yet metric 3D hand reconstruction from this very signal still has neither an end-to-end model nor an in-the-wild benchmark. We propose ESTHER, a model whose stereo geometry, temporal reasoning, and output representation are designed for wearable egocentric stereo. It is trained on pseudo-labels from a calibrated labeling pipeline and in turn assembles our benchmark ESTHER3D, an egocentric stereo hand dataset pairing a large in-the-wild training set of model-generated labels with a motion capture test set of true metric ground truth. Experiments show state-of-the-art accu?racy, superior external generalization, and robustness to the missing views, dropped frames, and lighting and motion blur extremes of real egocentric capture that break existing meth?ods. This robustness runs deeper than graceful degradation: stereo guidance teaches the model to bind apparent hand scale to metric depth, so it not only adapts to different stereo rigs and modalities with minimal fine-tuning, but more strikingly preserves true metric scale even after collapsing to a single monocular view.
| Subjects: | Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI) |
| Cite as: | arXiv:2609.34817 [cs.CV] |
| (or arXiv:2609.34817v2 [cs.CV] for this version) | |
| https://doi.org/10.48550/arXiv.2609.34817 arXiv-issued DOI via DataCite |
Submission history
From: Hongyu Ma [view email]
[v1]
Mon, 28 Sep 2026 10:17:22 UTC (26,627 KB)
[v2]
Sat, 10 Oct 2026 12:53:12 UTC (31,876 KB)
来源：arXiv：cs.AI · arxiv.org