Abstract:3D face alignment of monocular images is a crucial process in the recognition of faces with disguise.3D face reconstruction facilitated by alignment can restore the face structure which is helpful in detcting disguise <a class="link-external link-http" href="http://interference.This" rel="external noopener nofollow">this http URL</a> paper proposes a dual attention mechanism and an efficient end-to-end 3D face alignment <a class="link-external link-http" href="http://framework.We" rel="external noopener nofollow">this http URL</a> build a stable network model through Depthwise Separable Convolution, Densely Connected Convolutional and Lightweight Channel Attention Mechanism. In order to enhance the ability of the network model to extract the spatial features of the face region, we adopt Spatial Group-wise Feature enhancement module to improve the representation ability of the network. Different loss functions are applied jointly to constrain the 3D parameters of a 3D Morphable Model (3DMM) and its 3D vertices. We use a variety of data enhancement methods and generate large virtual pose face data sets to solve the data imbalance problem. The experiments on the challenging AFLW,AFLW2000-3D datasets show that our algorithm significantly improves the accuracy of 3D face alignment. Our experiments using the field DFW dataset show that DAMDNet exhibits excellent performance in the 3D alignment and reconstruction of challenging disguised <a class="link-external link-http" href="http://faces.The" rel="external noopener nofollow">this http URL</a> model parameters and the complexity of the proposed method are also reduced <a class="link-external link-http" href="http://significantly.The" rel="external noopener nofollow">this http URL</a> code is publicly available at https:// <a class="link-external link-http" href="http://github.com/LeiJiangJNU/DAMDNet" rel="external noopener nofollow">this http URL</a>

Multi-Attention Network for 2D Face Alignment in the Wild.

Multi‐scale Cross‐domain Alignment for Person Image Generation

Selective Domain-Invariant Feature Alignment Network for Face Anti-Spoofing.

Real-Time Facial Landmark Detection by Attention-driven Lightweight Network

SADRNet: Self-Aligned Dual Face Regression Networks for Robust 3D Dense Face Alignment and Reconstruction

MTCNN and FACENET Based Access Control System for Face Detection and Recognition

Multi-task Convolution Network for Face Alignment

Distract Your Attention: Multi-Head Cross Attention Network for Facial Expression Recognition

A multi-branch attention and alignment network for person re-identification

Multi-face detection and alignment using multiple kernels

Dual Attention MobDenseNet(DAMDNet) for Robust 3D Face Alignment

An Efficient Multitask Neural Network for Face Alignment, Head Pose Estimation and Face Tracking

Multistage Model for Robust Face Alignment Using Deep Neural Networks

Joint Multi-View Face Alignment in the Wild

Robust 3D Face Alignment with Multi-Path Neural Architecture Search

Cascaded Deep Convolutional Neural Network for Robust Face Alignment.

Multiscale recurrent regression networks for face alignment

Concentrated Multi-Grained Multi-Attention Network for Video Based Person Re-Identification

Contributions of glycolysis and oxidative phosphorylation to adenosine 5'-triphosphate production in AS-30D hepatoma cells.

Joint Face Detection and Alignment Using Multitask Cascaded Convolutional Networks

Cross Task Modality Alignment Network for Sketch Face Recognition