arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2603.02063cs.CV
收藏

ORGAN:基于循环一致生成对抗网络的对象中心表示学习

ORGAN: Object-Centric Representation Learning using Cycle Consistent Generative Adversarial Networks

Laboratory of Biosensors and Bioelectronics, Institute for Biomedical Engineering, University and ETH Zurich(生物传感器与生物电子实验室,生物医学工程研究所,大学和ETH苏黎世分校);Department of Neurology, Insel Gruppe, Bern, Switzerland(神经病学系,因塞格鲁普,瑞士伯恩);University of California, San Francisco, USA(加州大学旧金山分校,美国);Department of Neurobiology, University of Chicago, 951 E 58th St, Chicago, 60637, IL, USA(神经生物学系,芝加哥大学,951 E 58th St, 奇克阿哥,60637, IL, 美国);Department of Physics, University of Chicago, 929 E 57th St, Chicago, 60637, IL, USA(物理学系,芝加哥大学,929 E 57th St, 奇克阿哥,60637, IL, 美国)

Joël Küchler, Ellen van Maren, Vaiva Vasiliauskaitė, Katarina Vulić, Reza Abbasi-Asl, Stephan J. Ihle

AI总结:

ORGAN通过循环一致生成对抗网络实现对象中心表示学习,在合成数据集上表现优异,并能处理现实世界中具有众多对象和低对比度的数据。

摘要:

尽管数据生成通常较为简单,但从数据中提取信息更具挑战性。对象中心表示学习可以通过将图像分割为子组件(即对象)来以无监督的方式提取信息。每个对象在低维潜在空间中进行表示,可用于后续处理。对象中心表示学习主要由自编码器架构(AEs)主导。在这里,我们提出了ORGAN,一种新的对象中心表示学习方法,其基于循环一致的生成对抗网络。我们展示了ORGAN在合成数据集上与其它最先进的方法表现相似,同时是唯一一个在此测试中能够处理具有许多对象和低视觉对比度的更具挑战性的现实世界数据集的方法。补充这些结果,ORGAN创建了具有表达力的潜在空间表示,允许对象操作。最后,我们展示了ORGAN在对象数量和图像大小方面具有良好的可扩展性,使其在当前最先进的方法上具有独特的优势。

英文摘要:

Although data generation is often straightforward, extracting information from data is more difficult. Object-centric representation learning can extract information from images in an unsupervised manner. It does so by segmenting an image into its subcomponents: the objects. Each object is then represented in a low-dimensional latent space that can be used for downstream processing. Object-centric representation learning is dominated by autoencoder architectures (AEs). Here, we present ORGAN, a novel approach for object-centric representation learning, which is based on cycle-consistent Generative Adversarial Networks instead. We show that it performs similarly to other state-of-the-art approaches on synthetic datasets, while at the same time being the only approach tested here capable of handling more challenging real-world datasets with many objects and low visual contrast. Complementing these results, ORGAN creates expressive latent space representations that allow for object manipulation. Finally, we show that ORGAN scales well both with respect to the number of objects and the size of the images, giving it a unique edge over current state-of-the-art approaches.

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。

↑