Improving Face Detection Performance with 3D-Rendered Synthetic Data
This work addresses the challenge of data scarcity and variability in face detection for computer vision applications, but it is incremental as it builds on existing synthetic data methods.
The paper tackled the problem of improving face detection by using 3D-rendered synthetic data with controlled variations, and found that this approach enhanced performance across multiple detectors and datasets, though specific numbers were not provided.
In this paper, we provide a synthetic data generator methodology with fully controlled, multifaceted variations based on a new 3D face dataset (3DU-Face). We customized synthetic datasets to address specific types of variations (scale, pose, occlusion, blur, etc.), and systematically investigate the influence of different variations on face detection performances. We examine whether and how these factors contribute to better face detection performances. We validate our synthetic data augmentation for different face detectors (Faster RCNN, SSH and HR) on various face datasets (MAFA, UFDD and Wider Face).