Master Assignment
TURBULENCE: 2D gen – 3D pose pipeline – atmospheric turbulence
Type: Master EE/CS
Period: TBD
Student: (Unassigned)
If you are interested please contact :
Description:
One of the major challenges in face recognition at a distance (medium roughly 300m, long roughly 1000m) is the presence of atmospheric turbulence, causing temporal and spatial changes to the image extracted from the image sensor.
Generative networks are powerful structures that can create very natural looking facial images, see for example https://thispersondoesnotexist.com/. Also, it is possible to create a 3D face model based on a 2D image.
The purpose of this assignment is to create a realistic simulation of the effect of atmospheric turbulence, not only for frontal faces but also for facial images taken from a non-frontal, elevation angle perspective.
Assignment
- Create 2D-3D-atmosperic turbulence pipeline.
- Compare results with results found in literature.