Je suis intéressé
The project aims at efficient deep learning algorithms for facial recognition from standard camera images. It also involves embedding these algorithms in a miniature vision system with a low-power off-the-shelf processor. Facial analytics and recognition is a rapidly growing market, with applications ranging from surveillance and access control systems in smart buildings to retail stores that collect viewership and demographics. In application areas with low-power, small-size and low-cost requirements, a major barrier to rapid adaptation of the technology is the computationally demanding nature of the algorithms for high reliability. The project’s objective is to address this barrier by developing efficient deep learning architectures that can be run in real time on low-cost hardware platforms with limited computational resources.
To understand the existing deep learning architectures used at CSEM, redesign them for improved efficiency and to embed them in a low-power processor, potentially a hardware accelerator. The challenge here will be to find an efficient architecture for the facial recognition problem and optimize it to maintain reliability in recognition accuracy while minimizing the computational resources required to evaluate the AI model. The optimized model will be embedded in a camera system to demonstrate its efficiency and recognition performance. The candidate is expected to present their results at the end of the project, which will last between 4-6 months.
You have a background in machine learning, especially deep learning techniques, computer vision, and programming in C/C++/python. You are interested in understanding as well as designing intelligent algorithms and are a good team player.
CSEM offers a stimulating and multidisciplinary work environment with the opportunity to work with leading Swiss and international companies. You will have the opportunity to benefit from a multicultural company which clearly promotes an employee-driven culture. CSEM is an equal opportunities employer.
For further information on this position, please contact Dr. Engin Türetken, Senior R&D Engineer in the Vision Embedded Systems group (+4132 720 5237, firstname.lastname@example.org).
We look forward to receiving your complete application file at email@example.com, mentioning ref. M111.2018-3 in the subject. Preference will be given to professionals applying directly.