研究人员开发了一种新的训练流程，以解决结合了图像和文本数据的行人重识别（ReID）系统中的优化冲突。所提出的解耦两阶段方法使用单一视觉编码器来学习共享表示，而不会受到基于图像（I2I）和基于文本（T2I）的检索目标之间的干扰。实验表明，使用I2I ReID进行预训练可以提高对T2I任务的泛化能力，并且在视觉编码器训练期间加入文本监督可以提高两种模态的性能。 AI
影响 这项研究提供了一种提高多模态人员识别系统的准确性和泛化能力的方法。
排序理由 该集群包含一篇学术论文，详细介绍了特定AI任务的新方法。
在 Hugging Face Daily Papers 阅读 →
AI 生成摘要 · Google Gemini · 来自 3 个来源。 我们如何撰写摘要 →
完整方法见我们的编辑标准。
arXiv:2606.02242v1 Announce Type: cross Abstract: The joint optimization of image-based (I2I) and text-based (T2I) person re-identification (ReID) is hindered by modality discrepancies and conflicting training objectives, leading to suboptimal shared representations. While I2I Re…
The joint optimization of image-based (I2I) and text-based (T2I) person re-identification (ReID) is hindered by modality discrepancies and conflicting training objectives, leading to suboptimal shared representations. While I2I ReID focuses on identity-level invariance across ima…
The joint optimization of image-based (I2I) and text-based (T2I) person re-identification (ReID) is hindered by modality discrepancies and conflicting training objectives, leading to suboptimal shared representations. While I2I ReID focuses on identity-level invariance across ima…