[robotics-worldwide] [Jobs] PhD position Building Editable 3D Models using Deep Learning

Patric Jensfelt patric at kth.se
Thu Jul 18 04:23:38 PDT 2019

In this doctoral project, fully funded by the national research program WASP (wasp-sweden.se<https://urldefense.proofpoint.com/v2/url?u=http-3A__wasp-2Dsweden.se&d=DwIFAg&c=clK7kQUTWtAVEOVIgvi0NU5BOUHhpN0H8p7CSfnc_gI&r=0w3solp5fswiyWF2RL6rSs8MCeFamFEPafDTOhgTfYI&m=Lq-_bnDN78Ar8fOWy8Dl0AEtRkmDOB_r0lkT4rgNbtY&s=6GYIcLmxTCsMXhwqLt-Pt6vHMNZ5HalGZQGE1aofFZk&e= >), we will study the problem of autonomously generating, refining and manipulating dense 3D models. We will combine machine learning techniques for semantic and object level perception with traditional sensor-based robot mapping. The distinguishing feature of this project is that we will treat geometry and semantics jointly to form an overall more accurate and complete map and to allow the model to be edited interactively.  The vision of the project is a mapping framework that allows for autonomous creation of 3D models that support, for example, VR type applications. We will take a combined bottom-up and top-down approach. The starting point will be to generate a dense 3D model of the scene using real sensor data.

You would be employed in the Division for Robotics, Perception, and Learning (RPL) at the School of Electrical Engineering and Computer Science at KTH Royal Institute of Technology in Stockholm Sweden. RPL currently has 13 faculty and about 55 doctoral students covering a wide range of topics, providing a good environment for the research.

Deadline: July 31, 2019

Apply here

Patric Jensfelt

More information about the robotics-worldwide mailing list