Presentation of our new paper at the Cyberworlds conference
Former student Christon Ragavan Nadar will present the paper Sensor Simulation for Monocular Depth Estimation using Deep Neural Networks at Cyberworlds 2021. The underlying work was done in a media project supervised by Christian Kunert in the context of the TAB project ARCHITECT. The paper is about the estimation of depth maps based on simple color images taken with an iPad. For this purpose, existing datasets and neural networks are used, which are re-trained using a transfer learning approach and adapted in such a way that they can reproduce depth sensors for the iPad in a detailed manner. The conference will take place in Caen from September 28-30, and due to the current pandemic, it is a hybrid event that will take place online in parallel. Our paper will be presented on 9/20 from 10:00 to 10:20.