PulseAugur
实时 17:23:01

REViT: Vision Transformer 实现旋转反射等变性

研究人员开发了 REViT,一种新颖的视觉 Transformer,它使用卷积注意力来整合旋转反射等变性。这种方法旨在保留特征图中的旋转和翻转对称性,这对于图像分类和物体检测等输入方向至关重要的任务特别有益。该研究解决了在视觉 Transformer 中实现等变性的挑战,并提出了一种简化的方法,据报道,该方法在图像分类的离散旋转反射群等变神经网络方面优于现有技术。 AI

影响 这项研究可能带来更强大的计算机视觉模型,能够更好地处理图像中的方向变化。

排序理由 该集群包含一篇描述新模型架构的学术论文。[lever_c_demoted from research: ic=1 ai=1.0]

在 arXiv cs.LG 阅读 →

AI 生成摘要 · Google Gemini · 来自 3 个来源。 我们如何撰写摘要 →

REViT: Vision Transformer 实现旋转反射等变性

报道来源 [3]

  1. arXiv cs.LG TIER_1 English(EN) · Sheir A. Zaheer, Alexander C. Holston, Chan Y. Park ·

    REViT: Roto-reflection Equivariant Convolutional Vision Transformer

    arXiv:2606.25318v1 Announce Type: cross Abstract: In this paper, we propose a discrete roto-reflection group equivariant vision transformer with convolutional attention. Roto-reflection equivariant networks preserve the rotational, flip and positional symmetry in feature maps, ma…

  2. arXiv cs.LG TIER_1 English(EN) · Chan Y. Park ·

    REViT: Roto-reflection Equivariant Convolutional Vision Transformer

    In this paper, we propose a discrete roto-reflection group equivariant vision transformer with convolutional attention. Roto-reflection equivariant networks preserve the rotational, flip and positional symmetry in feature maps, making them useful for tasks where orientation of th…

  3. Hugging Face Daily Papers TIER_1 English(EN) ·

    REViT: Roto-reflection Equivariant Convolutional Vision Transformer

    In this paper, we propose a discrete roto-reflection group equivariant vision transformer with convolutional attention. Roto-reflection equivariant networks preserve the rotational, flip and positional symmetry in feature maps, making them useful for tasks where orientation of th…