Adaptive CMOS Image Sensor for smart vision systems
Published : 10 January 2019
The aim of this thesis is to explore new kind of smart vision sensor architectures using for enhance the sensor reactivity and for simplify the image processing. The studied vision system will use new 3D microelectronic technologies from CEA-leti. These technologies are capable to stack several integrated circuits. The main advantage is to propose a high density of interconnections between them, allowing connection at the pixel level. This characteristic allows us to think about a totally new architecture of the image processing chain of a basic imager (readout, amplification, compensation, colorization, tone mapping) in order to improve the agility, a better image quality, a better energy efficiency, with a low silicon footprint.
The PhD student will benefit during his 3-years thesis of the expertise and the scientific excellence of the CEA leti to attend objectives with a high level of innovation through international patents and publications.
The dynamic and autonomous candidate, will have a microelectronic master degree, specialized in analog integrated circuit design. A good knowledge of circuit design CAD tools will be important (Cadence, and also Matlab) and good knowledge in image processing will be appreciated.
This thesis will start with the state of the art study, then the PhD student will define the optimal architecture. Finally, a test chip will be designed and tested. It will demonstrate the scientific and industrial potentialities of the proposed solutions.