Abstract
In current practical face authentication systems, most face recognition (FR) algorithms are based on cosine similarity with softmax classification. Despite its reliable classification performance, this method struggles with hard samples. A popular strategy to improve FR performance is incorporating angular or cosine margins. However, it does not take face quality or recognition hardness into account, simply increasing the margin value and thus causing an overly uniform training strategy. To address this problem, a novel loss function is proposed, named Loss function for Hard High-quality Face (LH2Face). Firstly, a similarity measure based on the von Mises-Fisher (vMF) distribution is stated, specifically focusing on the logarithm of the Probability Density Function (PDF), which represents the distance between a probability distribution and a vector. Then, an adaptive margin-based multi-classification method using softmax, called the Uncertainty-Aware Margin Function, is implemented in the article. Furthermore, proxy-based loss functions are used to apply extra constraints between the proxy and sample to optimize their representation space distribution. Finally, a renderer is constructed that optimizes FR through face reconstruction and vice versa. Our LH2Face is superior to similiar schemes on hard high-quality face datasets, achieving 49.39% accuracy on the IJB-B dataset, which surpasses the second-place method by 2.37%.
Abstract (translated)
在当前的实际面部认证系统中，大多数面部识别（FR）算法都是基于余弦相似度和softmax分类。尽管这种方法具有可靠的分类性能，但在处理难以识别的样本时表现不佳。一种常用的改进面部识别性能的方法是引入角度或余弦边距。然而，这种策略没有考虑面部质量或识别难度的问题，只是简单地增加边距值，从而导致了过于统一的训练策略。为了解决这个问题，提出了一种新的损失函数，名为“针对硬高质量人脸的损失函数”（Loss function for Hard High-quality Face, LH2Face）。 首先，该方法基于冯·米塞斯-费希尔(von Mises-Fisher, vMF)分布提出了一个相似度测量标准，并特别关注了概率密度函数(PDF)的对数值，这代表了一个概率分布与向量之间的距离。然后，在文章中实现了一种使用softmax的不确定性感知边距函数的多分类方法，这是一种自适应边距基的方法。此外，还采用了基于代理的损失函数来在代理和样本之间施加额外约束，以优化它们的表示空间分布。最后，构建了一个渲染器，通过面部重建和反向过程来进行面部识别的优化。 我们的LH2Face方案在硬高质量的人脸数据集上优于类似方案，在IJB-B数据集上的准确率为49.39%，比第二名的方法高出2.37%。
URL
https://arxiv.org/abs/2506.23555
