Fraunhofer-Institut für Integrierte Schaltungen IIS

Interns for Deep Learning for Audio-Visual Speech Separation

Online seit
14.08.2019 - 14:50
Job-ID
ERN-2019-08-14-269251
Kategorie
Praktika

Job-Beschreibung

For the Audio and Media Technologies division in Erlangen, the Fraunhofer Institute for Integrated Circuits IIS is currently seeking

Master Thesis Students or Interns for Deep Learning for Audio-Visual Speech Separation

The »Audio and Media Technologies« division of Fraunhofer IIS constitutes one of the world’s largest organizations dedicated to audio, speech and media processing. It has been an innovator in sound and vision for over 25 years and repeatedly won international competitions in the audio and media field: e.g. with mp3 and codecs of the AAC-family. Over 200 engineers and scientists develop first-rate technology, which is sold worldwide, in Europe, USA, China, Korea and Japan.

What is this about?

The use of deep learning for solving complex tasks in audio and video signal processing has vastly increased. It is widely used e.g. in smart speakers and voice assistants, object recognition from images and videos in autonomous vehicles. Typically, machine learning methods are applied separately to either the recorded audio signal or the video. Recently, new concepts for the development of deep neural networks (DNN) for joint audio-visual signal processing have been proposed to extract speech signals from complex mixtures of sound, providing improved performance compared to processing each signal modality separately.

Your responsibilities:

Your first task would be to get an overview of current state-of-the-art for joint audio-visual processing with deep learning. A DNN model for speech signal separation will be developed using common deep learning frameworks (e.g. TensorFlow) and Python. This includes the implementation of a complete workflow for training and testing of the DNN with publicly available datasets. Application-specific optimization with respect to signal quality and efficiency of the developed baseline DNN model will conclude your work.

Anforderungsprofil

What we expect from you: You …

  • are interested or skilled in deep learning.
  • have experience with Python
  • have basic knowledge of digital signal processing.
  • are motivated to learn and work on complex tasks within a team of researchers

What you can expect from us

  • An open and cooperative working environment
  • Collaboration in interesting and innovative projects
  • Many opportunities to gain practical experience
  • Flexibility concerning your working hours

If you have any questions about this opening, please contact soumitro.chakrabarty@iis.fraunhofer.de

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.

Interested?

Please apply for this position using the following link: https://recruiting.fraunhofer.de/Vacancies/47295/Description/2

Applications are possible in German and English. Please include a cover letter, your CV and your latest transcripts of records (as PDF) and quote ID number 47295. Address your application to Nina Wörlein.

Please let us know how you learned about this job opportunity.

Additional information is available on our website: www.iis.fraunhofer.de/en

Firmen- und Kontaktdaten

Firmenname
Fraunhofer-Institut für Integrierte Schaltungen IIS
Ansprechpartner
Frau Nina Wörlein
Einsatzort
Am Wolfsmantel
33
91058 Erlangen
Deutschland
Telefon
+49 9131 7761684
E-Mail
Webseite

Art der Beschäftigung

nach Vereinbarung

Weitere Informationen

Zeitraum der Beschäftigung
nach Vereinbarung