English(EN) Low-Quality Face Recognition using Center Aligned Representations and Local Margin Constraints

新框架提升低质量人脸识别准确率

作者 PulseAugur 编辑部 · [1 个来源] · 2026-09-02 04:00

研究人员开发了一个新框架，以提高低质量图像的人脸识别准确率。该框架通过引入局部概率边界（LPM）来估计样本难度，使用嵌套注意力模块（NAM）增强Transformer层，并采用质量门控协议（QGP）根据图像质量调整适配器贡献，从而解决了匹配退化图像的挑战。在TinyFace、SurvFace、IJB-B和IJB-C等基准测试上的实验表明，在识别和验证任务上均取得了显著改进。 AI

影响提高了在图像质量下降情况下运行的人脸识别系统的准确性。

排序理由该条目是一篇学术论文，详细介绍了一种用于特定计算机视觉任务的新方法。[lever_c_demoted from research: ic=1 ai=1.0]

在 arXiv cs.CV 阅读 →

AI 生成摘要 · Google Gemini · 来自 1 个来源。我们如何撰写摘要 →

本文如何被排名

Signal score: 25 / 100; Composite score across the factors below. Higher = stronger signal that this story matters right now.
Newsworthiness bucket: Tool; 该条目是一篇学术论文，详细介绍了一种用于特定计算机视觉任务的新方法。[lever_c_demoted from research: ic=1 ai=1.0]
Source corroboration: Single-source cluster; Only one publisher covered this so far. Single-source stories can still rank when the publisher is high-authority, but they lack cross-source corroboration.
Topics: paper, other; Editorial topic classification. Feeds into how the story surfaces on /topic/<slug> hub pages and into the per-entity coverage mix.
AI-industry relevance: High; Clearly on-topic for AI-industry coverage.
Story freshness: Breaking (< 6h); Fresh story with cross-source coverage still developing. Ranking may shift as more sources report.

完整方法见我们的编辑标准。

报道来源 [1]

arXiv cs.CV TIER_1 English(EN) · Vedat Can Dilaver, Benjamin S. Riggan · 2026-09-02 04:00

使用中心对齐表示和局部边距约束的低质量人脸识别

arXiv:2609.01014v1 Announce Type: new Abstract: Low-quality face recognition (LQFR) remains challenging due to the difficulty of matching degraded query (probe) images against low-quality (LQ) enrollment (gallery) imagery and the scarcity of training data for large-scale models. …

本文如何被排名

报道来源 [1]

使用中心对齐表示和局部边距约束的低质量人脸识别

相关话题