Master Electrical Engineering

Develop a model-based approach for 3D object recognition with Microsoft HoloLens

Develop a model-based approach for 3D object recognition with Microsoft HoloLens

Student : unassigned
Type : M-EE
Duration : November 2016 until unknown
Supervisors : Dr.ir. L.J. Spreeuwers (Luuk), Prof.Dr.Ir. R.N.J. Veldhuis (Raymond)
Faculty : EEMCS

Background
This assignment is from a startup founded by three EWI graduates, that is currently developing its technology under the radar. We are building a developer platform for Augmented Reality; providing spatial understanding of user's environments in order to create truly immersive experiences. Our first product applies machine learning to our data set to help developers interact with their user’s environment. We are a remote company with team-members based in The Netherlands and San Francisco Bay Area.

The assignment
Right now, we apply a machine learning based approach to translate our data set into recognized objects. The assignment goal is to define a model-based approach that improves the quality and efficiency of the object recognition. How a model-based approach can improve our object recognition is part of the research itself. Key component of the assignment is to define how and where in our data processing cycle a model-based approach fits best. This can be by working alongside machine learning as separate track, using the machine learning output as ongoing starting point, or even utilizing output from the machine learning to develop a model-based approach that has the potential to replace the machine learning approach altogether. Key part of the assignment is to develop a first model-based approach and execute real-life testing based on our constantly growing dataset from HoloLens developers.
Final deliverables of the assignment are:
1.Conclusion if a model-based approach can improve the quality and efficiency of our object recognition.
2.Recommendation where in our data processing cycle this approach should sit, together with or even replacing machine learning.
3.Working version of a model-based approach, together with testing results benchmarked with the current machine learning output.

About you
You are a graduation student that has a passion for 3D and get’s really excited by working with new Augmented Reality technologies like HoloLens. You get to work in the core of a very driven team and this is a fundamental piece of our technological challenge.