arXiv cs.AI· Despoina Touska, Bastiaan Onne Fagginger Auer, Alexandru Onose, Tejaswi Kasarla, Luis Armando P\'erez Rey, Maximilian Lipp, Lyubov Amitonova, Martin R. Oswald, Pascal Cerfontaine·· 3 小时前AI 评分35
OrthoRF：探索对象中心表示中的正交性
OrthoRF: Exploring Orthogonality in Object-Centric Representations
AI 导读
研究者提出 Orthogonal Rotating Features（OrthoRF），通过内积损失和架构修改在 Rotating Features 的取向空间中强制正交性，从而获得更锐利的相位对齐和更可靠的对象分组。在无监督对象发现评测（含重叠对象、噪声和分布外测试）中，OrthoRF 匹配或超越现有模型，并免除了许多基于同步性方法所需的事后聚类，还能恢复被遮挡对象的部分。
正文 · 原文
Abstract:Neural synchrony is hypothesized to help the brain organize visual scenes into structured multi-object representations. In machine learning, synchrony-based models analogously learn object-centric representations by storing binding in the phase of complex-valued features. Rotating Features (RF) instantiate this idea with vector-valued activations, encoding object presence in magnitudes and affiliation in orientations. We propose Orthogonal Rotating Features (OrthoRF), which enforces orthogonality in RF's orientation space via an inner-product loss and architectural modifications. This yields sharper phase alignment and more reliable grouping. In evaluations of unsupervised object discovery, including settings with overlapping objects, noise, and out-of-distribution tests, OrthoRF matches or outperforms current models while producing more interpretable representations, and it eliminates the post-hoc clustering required by many synchrony-based approaches. Unlike current models, OrthoRF also recovers occluded object parts, indicating stronger grouping under occlusion. Overall, orthogonality emerges as a simple, effective inductive bias for synchrony-based object-centric learning.
| Comments: | Poster presentation at ICLR 2026 |
| Subjects: | Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI); Machine Learning (cs.LG) |
| Cite as: | arXiv:2610.35203 [cs.CV] |
| (or arXiv:2610.35203v1 [cs.CV] for this version) | |
| https://doi.org/10.48550/arXiv.2610.35203 arXiv-issued DOI via DataCite (pending registration) |
Submission history
From: Despoina Touska [view email]
[v1]
Mon, 26 Oct 2026 10:55:14 UTC (4,059 KB)
来源：arXiv cs.AI · arxiv.org