Section: New Results

Face Analysis in Structured Light Images

Participants : Vikas Thamizharasan, Antitza Dantcheva, Francois Brémond.

Keywords: Structured light, Face analysis

The main objective has been to perform face analysis tasks like authentication, gender, age and ethnicity classification by generating low-dimensional face embedding from the raw data acquired from structured light (see Figure 14) sensors using deep learning techniques. In this context we studied depth/disparity map extraction (see Figure 15), as well as other models.

Figure 14. Structured light. A calibrated camera and projector (typically both near infrared) are placed at a fixed, known baseline. The structured light pattern helps establish correspondence between observed and projected pixels. Depth is derived for each corresponding pixel through triangulation. The process is akin to two stereo cameras, but with the projector system replacing the second camera, and aiding the correspondence problem.

Figure 15. *IR - Infrared image, IRB - Binarized Infrared image