CSANet: Channel Spatial Attention Network for Robust 3D Face Alignment and Reconstruction

TL;DR

CSANet is a novel end-to-end 3D face alignment and reconstruction network that leverages advanced attention mechanisms and specialized loss functions to improve accuracy and training stability.

Contribution

It introduces a new architecture combining channel spatial attention with a robust training strategy for 3D face modeling.

Findings

Outperforms baseline models quantitatively

Achieves superior qualitative reconstruction results

Demonstrates stable training and convergence

Abstract

Our project proposes an end-to-end 3D face alignment and reconstruction network. The backbone of our model is built by Bottle-Neck structure via Depth-wise Separable Convolution. We integrate Coordinate Attention mechanism and Spatial Group-wise Enhancement to extract more representative features. For more stable training process and better convergence, we jointly use Wing loss and the Weighted Parameter Distance Cost to learn parameters for 3D Morphable model and 3D vertices. Our proposed model outperforms all baseline models both quantitatively and qualitatively.