For the conversion of sparse lightfields into dense ones, amongst others, accurate disparity maps are required. Obtaining these disparity maps for sparse lightfields isn’t as simple or straightforward as one would expect. Due to the wider baseline, occlusions and lighting conditions have a bigger impact. New algorithms need to be developed or existing ones need to be adapted to be able to cope with this. Apart from these added challenges, well-known challenges like homogeneous and reflective surfaces also need to be handled. We focus on Lambertian scenes.
So far existing SOTA disparity estimation methods have been studied and evaluated and new ideas have been implemented and evaluated. A next step would be to look into more details regarding the view rendering step, which generates the denser lightfield.
I’ve attended the third ETN-FPI training school in Kiel and in November I will start my first secondment at the CAU in Kiel.
Ron holds Bachelor’s and Master’s degrees in Electrical Engineering from the University of Technology Eindhoven (Netherlands). In his Master’s thesis, he focused on the handling of occlusions in an object detection system using computer vision and machine learning.
Moving Picture Technologies
Fraunhofer-Institut für Integrierte Schaltungen IIS
Am Wolfsmantel 33