Doctoral Researcher, Signal Processing and Machine Learning
Open position
In collaboration with
About
The Doctoral School of Industry Innovations (DSII) with Tampere University (TAU) explores research questions that are directly relevant to industry and business.
Tampere University hires a Doctoral Researcher to complete a four-year dissertation project under the joint supervision of two university professors and two company representatives. The position is placed in the 3D Media Group, which is part of the Computing Sciences Unit within the Faculty of Information Technology and Communication Sciences (ITC).
DSII offers Doctoral Researchers a unique opportunity to undertake dissertation research, gain a thorough understanding of product development, tackle real-world business challenges and expand their professional networks.
The partner company for this position is Microsoft, through its Surface Imaging Team in Tampere, which researches and develops new technologies in the fields of imaging, AI, mobile devices, and electronics.
Job description
The purpose of the job is to create new knowledge for academia and industry in the fields of computational imaging, optics and machine learning. The research project aims at jointly optimizing the optics and inverse imaging algorithms for compact camera designs employing machine learning to achieve image quality for a certain task.
The research goals are specified as follows:
1. Develop models of refractive optics applicable to compact optics used in portable devices that ensure realistic and manufacturable compact optics design
2. Develop ML based post-processing imaging methods to couple with the optical models
3. Develop end-to-end imaging system optimization methods to co-design compact optics and efficient ML algorithms
4. Validate the end-to-end methodology on selected case(s) of image restoration and enhancement
As a Doctoral Researcher you are expected to perform overviews of state of the art, to study the principles of physical image formation and its modelling in terms of optics, sensing and image processing, to investigate methods for optical/imaging co-design employing modern machine learning an develop the corresponding algorithms, to generate experimental results, analyse and validate them on use case(s) through comparison with the state of the art.
By participating in DSII events, excursions, meetings and training sessions with your fellow doctoral students, you will both get and provide peer support and make your own contribution to the development of DSII.
Requirements
You need to have an applicable master’s degree in signal processing, electronics, physics, optics and photonics, information technology, computer science, or a related discipline in science or engineering.
The position requires:
1. Strong background at MSc level in:
- Mathematics: linear algebra, numerical methods, probability and statistics
- Signal and image processing: sampling and reconstruction of signals, Fourier transform, image formation and image processing
- Optics: basics of geometrical and wave optics
-2. Good command of both written and spoken English
Experience in some of the following fields is considered as a plus:
1. Vector-space formalism
2. Constrained and unconstrained optimization
3. Optical design with Zemax or similar software
4. Machine learning
- Neural networks: DNN, CNN, RNN, Transformers
- Learning: Loss functions, differentiation and. Back propagation, stochastic gradient descent, regularization
- Deep Generative models: plug-and-play models, variational autoencoders, GANs
Please note that before commencing in the position, the selected candidate must apply for a study right within the doctoral programme at Tampere University, unless they already have one. Please visit the admissions webpage for more information on eligibility requirements.
Contact
Prof. Atanas Gotchev