View job here

Smart speakers should be able to adapt to their environment for optimizing sound reproduction [1]. To achieve this, it is necessary to estimate the geometry of the room in which they are placed. There are two main approaches to accomplishing this: The first approach is through acoustic methods. These methods involve sending out sound waves from the smart speaker and measuring the time it takes for the waves to return. By analyzing this data, the geometry of the room can be derived. Alternatively, geometry information can also be obtained through the use of images taken from different perspectives. This methods rely on fundamental principles of structure for motion, allowing for the derivation of a 3D model.

To contribute to the improvement of future smart speaker technology, your task is to compare these two modalities and identify their respective strengths and weaknesses. This analysis will serve as a foundation for finding clever ways to combine them in order to achieve higher quality reconstruction results.

You are interested in improving the technology of future smart speakers?

Then have a look at our offer!

What you will do

  • You capture visual and acoustic data for several room types
  • You do the visual processing including camera calibration, voxel reconstruction [2], and mesh generation
  • You convert audio room geometry into a visual model
  • You align the visual and acoustic model for comparison
  • You evaluate strengths, weaknesses, precisions and completeness of the two modalities

What you bring to the table

  • You are currently studying computer science, computer graphics or signal processing.
  • You are interested in 3D modelling or 3D reconstruction
  • You have knowledge in Python, Pytorch and basic signal processing
  • Optional you have basic skills in using Blender

What you can expect

  • Flexible working hours
  • Open and friendly team work
  • Varied tasks with room for creativity
  • Exciting seminars and events
  • Networking with scientists
  • Active contribution in applied research
  • Interesting an innovative projects

Weekly working hours are determined by agreement. You can start from now on (as a student assistant from 10 to 20 hours a week or as an intern for a period of at least three months). You can reduce your hours before exams and increase them during semester breaks. You can flexibly determine the working days. After your studies, you have the option of working with us full or part time.

We would be happy to offer you the opportunity to write a bachelor's or master's thesis in cooperation with us in the above-mentioned subject area. The thesis will be assigned and carried out in accordance with the rules of your university. For this reason, please discuss the thesis with a professor who can advise you over the course of the project.

We value and promote the diversity of our employees' skills and therefore welcome all applications - regardless of age, gender, nationality, ethnic and social origin, religion, ideology, disability, sexual orientation and identity.

References:
[1]    C. Tuna et al., ‘3D Room Geometry Inference Using a Linear Loudspeaker Array and a Single Microphone’, IEEEACM Trans. Audio Speech Lang. Process., vol. 28, pp. 1729–1744, 2020, doi: 10.1109/TASLP.2020.2998299.
[2]    M. Tancik et al., ‘Nerfstudio: A Modular Framework for Neural Radiance Field Development’. arXiv, Feb. 08, 2023. doi: 10.48550/arXiv.2302.04264.


Interested? 
Apply online now (PDF: cover letter, CV, transcripts). We look forward to getting to know you! 

Fraunhofer-Institute for Integrated Circuits IIS
www.iis.fraunhofer.de/en

Application Deadline: none