Abstract
Facial landmark detection is an important task in computer vision with numerous applications, such as head pose estimation, expression analysis, face swapping, etc. Heatmap regression-based methods have been widely used to achieve state-of-the-art results in this task. These methods involve computing the argmax over the heatmaps to predict a landmark. Since argmax is not differentiable, these methods use a differentiable approximation, Soft-argmax, to enable end-to-end training on deep-nets. In this work, we revisit this long-standing choice of using Soft-argmax and demonstrate that it is not the only way to achieve strong performance. Instead, we propose an alternative training objective based on the classic structured prediction framework. Empirically, our method achieves state-of-the-art performance on three facial landmark benchmarks (WFLW, COFW, and 300W), converging 2.2x faster during training while maintaining better/competitive accuracy. Our code is available here: this https URL.
Abstract (translated)
面部标志检测是计算机视觉中的一个重要任务，具有广泛的应用，如头部姿态估计、表情分析和人脸替换等。基于热图回归的方法被广泛应用，并在这一任务中取得了最先进的成果。这些方法通常通过计算热图上的argmax来预测一个地标点。然而，由于argmax不是可微的，这些方法采用了可微分近似Soft-argmax，以便能够在深度网络上进行端到端训练。 在这项工作中，我们重新审视了长期以来使用Soft-argmax的选择，并证明它并非唯一实现高性能的方法。相反，我们提出了一种基于经典结构化预测框架的替代训练目标。在实验中，我们的方法在三个面部标志基准数据集（WFLW、COFW和300W）上达到了最先进的性能，在训练过程中收敛速度提高了2.2倍，同时保持了更好的/竞争性的准确度。 我们的代码在此处提供：[此URL]
URL
https://arxiv.org/abs/2508.14929
