首页/论文/2203.08049

arXiv 2203.080492022-03-18cs.CV

On Hyperbolic Embeddings in 2D Object Detection

Robert Bosch GmbH（罗伯特·博世有限公司）；University of Freiburg（弗莱堡大学）

Christopher Lang, Alexander Braun, Abhinav Valada

备注：: 14 pages, 5 figures

阅读 PDF ↗

arXiv 页面 ↗HTML 原文 ↗

英文摘要相关论文

英文摘要：

Object detection, for the most part, has been formulated in the euclidean space, where euclidean or spherical geodesic distances measure the similarity of an image region to an object class prototype. In this work, we study whether a hyperbolic geometry better matches the underlying structure of the object classification space. We incorporate a hyperbolic classifier in two-stage, keypoint-based, and transformer-based object detection architectures and evaluate them on large-scale, long-tailed, and zero-shot object detection benchmarks. In our extensive experimental evaluations, we observe categorical class hierarchies emerging in the structure of the classification space, resulting in lower classification errors and boosting the overall object detection performance.

机构、总结、中文摘要均由AI分析而来，可能存在错误，请以论文正文为准。

阅读原文

阅读 PDF ↗

arXiv 页面 ↗HTML 原文 ↗

本文内容

英文摘要相关论文

机构、总结、中文摘要均由AI分析而来，可能存在错误，请以论文正文为准。

继续探索

相关论文

Hyperbolic Image Segmentation

University of Amsterdam（阿姆斯特丹大学）；Leiden University（莱顿大学）；Vrije Universiteit Amsterdam（阿姆斯特丹自由大学）

For image segmentation, the current standard is to perform pixel-level optimization and inference in Euclidean output embedding spaces through linear hyperplanes. In this work, we show that hyperbolic manifolds provide a valuable alternative for image segmentation and propose a tractable formulation of hierarchical pixel-level classification in hyperbolic space. Hyperbolic Image Segmentation opens up new possibilities and practical benefits for segmentation, such as uncertainty estimation and boundary information for free, zero-label generalization, and increased performance in low-dimensional…

2022-03-11cs.CV

超几何空间中的语义对齐用于开放词汇语义分割

Semantic Alignment in Hyperbolic Space for Open-Vocabulary Semantic Segmentation

Fulbright University Vietnam（富布赖特大学越南分校）

本文提出HyRo框架，通过解耦超几何空间中的层次和语义对齐，提升开放词汇语义分割性能。

2026-05-09cs.CV

Skin Lesion Recognition with Class-Hierarchy Regularized Hyperbolic Embeddings

Monash University（莫纳什大学）；The University of Sydney（悉尼大学）；Kāhu；The University of Queensland（昆士兰大学）；Alfred Health（阿尔弗雷德医疗）

In practice, many medical datasets have an underlying taxonomy defined over the disease label space. However, existing classification algorithms for medical diagnoses often assume semantically independent labels. In this study, we aim to leverage class hierarchy with deep learning algorithms for more accurate and reliable skin lesion recognition. We propose a hyperbolic network to learn image embeddings and class prototypes jointly. The hyperbola provably provides a space for modeling hierarchical relations better than Euclidean geometry. Meanwhile, we restrict the distribution of hyperbolic p…

2022-09-13cs.CV

平衡双曲嵌入是天然的分布外检测器

Balanced Hyperbolic Embeddings Are Natural Out-of-Distribution Detectors

University of Amsterdam（阿姆斯特丹大学）

针对深度学习中的分布外样本检测问题，本文提出平衡双曲学习方法，通过联合优化层级失真与子层级平衡构建双曲类别嵌入作为分类原型，在多数据集和评分函数上优于现有方法及对比学习SOTA，还支持层级化分布外泛化。

2025-06-11cs.CVcs.LG

Hyp-UML：基于不确定性感知度量学习的双曲图像检索

Hyp-UML: Hyperbolic Image Retrieval with Uncertainty-aware Metric Learning

Inria（法国国家信息与自动化研究所）；E.m. Data Technology Ltd.（E.m.数据技术有限公司）；Xi’an Jiaotong University（西安交通大学）

本文提出Hyp-UML方法，通过结合双曲空间图像嵌入与不确定性感知度量学习，有效表示层次化数据结构，并提升图像检索性能，最终实现最先进结果。

2023-10-22cs.CV

少样本学习中的双曲嵌入与欧氏嵌入：同一枚硬币的两面

Hyperbolic vs Euclidean Embeddings in Few-Shot Learning: Two Sides of the Same Coin

Carnegie Mellon University（卡内基梅隆大学）；Institute for Systems and Robotics（系统与机器人研究所）；Instituto Superior Técnico（高等技术学院）

本文研究少样本分类中双曲嵌入与欧氏嵌入的对比，发现双曲嵌入在共同半径下表现最佳，但固定半径的欧氏编码器可取得更优性能。

2023-09-18cs.CVcs.LG

更多相关论文收起更多论文

Hyperbolic Image Embeddings

Moscow Institute of Numerical Mathematics of the Russian Academy of Sciences（俄罗斯科学院莫斯科数值数学研究所）；Samsung AI Center Moscow（三星莫斯科人工智能中心）；Yandex（扬德克斯）

Computer vision tasks such as image classification, image retrieval and few-shot learning are currently dominated by Euclidean and spherical embeddings, so that the final decisions about class belongings or the degree of similarity are made using linear hyperplanes, Euclidean distances, or spherical geodesic distances (cosine similarity). In this work, we demonstrate that in many practical scenarios hyperbolic embeddings provide a better alternative.

2020-03-30cs.CVcs.LG

H²em：学习分层双曲嵌入用于组合零样本学习

$\text{H}^2$em: Learning Hierarchical Hyperbolic Embeddings for Compositional Zero-Shot Learning

HKUST（香港科技大学）；Zhejiang University（浙江大学）；ACCESS

H²em通过双曲几何学习分层嵌入，解决组合零样本学习中的层次结构建模问题，提升细粒度辨别能力和泛化能力。

2025-12-23cs.CV

OpenHype：用于层级开放词汇辐射场的双曲嵌入

OpenHype: Hyperbolic Embeddings for Hierarchical Open-Vocabulary Radiance Fields

TU Wien（维也纳技术大学）；Ulm University（乌尔姆大学）；Fondazione Bruno Kessler（布鲁诺·凯塞研究所）

OpenHype提出用连续双曲潜在空间表示3D场景层级，通过测地线路径实现多尺度关系编码与平滑遍历，在标准基准上超越现有方法，提升效率与适应性。

2025-10-24cs.CV

基于双曲嵌入的细粒度肋骨骨折诊断：详细标注框架与多标签分类模型

Fine-Grained Rib Fracture Diagnosis with Hyperbolic Embeddings: A Detailed Annotation Framework and Multi-Label Classification Model

Indian Institute of Technology Jodhpur（印度理工学院贾姆pur分校）；All India Institute of Medical Sciences Delhi（全印度医学科学研究所德里）；Government Medical College Srinagar（政府医学院斯里 Nagar）

针对现有肋骨骨折数据集缺乏细粒度标注的问题，提出新型标注方案与基于双曲嵌入的跨模态分类模型，在AirRib和RibFrac数据集上均取得更优的分类效果。

2025-04-16cs.CV

具有曲率意识优化的鲁棒双曲学习

Robust Hyperbolic Learning with Curvature-Aware Optimization

Data Science Group University of Hildesheim（数据科学组海德斯海姆大学）；ISMLL University of Hildesheim（ISMLL海德斯海姆大学）

本文提出一种鲁棒双曲学习方法，通过曲率意识优化提升泛化能力，减少计算成本和稳定性问题，提升多领域任务性能。

2026-02-07cs.CV

用于分层三维点云嵌入的双曲对比学习

Hyperbolic Contrastive Learning for Hierarchical 3D Point Cloud Embedding

East China Normal University（华东师范大学）

本文在双曲多模态对比预训练中引入三维点云模态，并设计蕴含、模态间隙与对齐正则化器学习模态内及跨模态层次嵌入，提升下游任务性能。

2025-01-07cs.AIcs.CV

Hyperbolic Vision Transformers: Combining Improvements in Metric Learning

University of Trento（特伦托大学）；Skolkovo Institute of Science and Technology（斯科尔科沃科技学院）；Skoltech（斯科尔科沃科技学院）；Yandex（杨德克斯）；AIRI（人工智能研究所）

Metric learning aims to learn a highly discriminative model encouraging the embeddings of similar classes to be close in the chosen metrics and pushed apart for dissimilar ones. The common recipe is to use an encoder to extract embeddings and a distance-based loss function to match the representations -- usually, the Euclidean distance is utilized. An emerging interest in learning hyperbolic data embeddings suggests that hyperbolic geometry can be beneficial for natural data. Following this line of work, we propose a new hyperbolic-based model for metric learning. At the core of our method is …

2022-03-22cs.CVcs.LG

HierLoc：基于超几何实体嵌入的层次视觉定位

HierLoc: Hyperbolic Entity Embeddings for Hierarchical Visual Geolocation

Huawei Riemann Lab（华为里曼实验室）；Hilbert Research Centre（希尔伯特研究中心）；Chair of Cartography, Technical University of Munich（慕尼黑技术大学制图学教授）

HierLoc通过超几何实体嵌入实现层次视觉定位，减少误差并提升细粒度准确率，为全球图像定位提供新方法。

2026-03-02cs.AIcs.CV

HYDEN：用于医学图像和报告的双曲密度表示

HYDEN: Hyperbolic Density Representations for Medical Images and Reports

Institute of Intelligent Diagnostics, Beijing United-Imaging Research Institute of Intelligent Imaging（北京联合智能影像研究院智能诊断研究所）；School of Physics, Peking University（北京大学物理学院）

针对医学图像与文本间的语义不确定性问题，提出HYDEN双曲密度嵌入方法，利用伪高斯分布映射特征并引入封装损失建模偏序关系，在零样本任务上优于基线。

2024-08-20cs.AIcs.CVcs.LG