CVIU_2

代码说明：

说明：  面部地标检测旨在定位给定特征点的面部图像的关键点，这些特征点通常遭受由任意姿势，不同面部表情和部分遮挡引起的变化。我们提出了一个粗到细的框架，它将堆叠的沙漏网络和显着区域的注意力细化联系起来，以实现稳健的面部对齐。为了实现这一目标，我们首先提出了一个多尺度区域学习模块（MSL）来分析不同面部区域的结构信息，并提取强烈的判别性深层特征。然后我们采用堆叠沙漏网络（SHN）进行热图回归和初始面部地标预测。(Facial landmark detection aims to locate keypoints for a facial image given feature points, which typically suffers from variations caused by arbitrary pose, diverse facial expressions, and partial occlusion. We put forward a coarse-to-fine framework which joints stacked hourglass network and salient region attention refinement for robust face alignment. In order to achieve this goal, we firstly put forward a multi-scale region learning module (MSL) to analyze the structure information at different facial region and extract strong discriminative deep feature. Then we employ stacked hourglass network (SHN) for heatmap regression and initial facial landmarks prediction.)