About us
Intelligent Embedded Vision is a Politehnica University Timișoara research group formed around a simple idea: intelligence is only useful once it runs where it's needed — on a chip inside a camera, a medical scanner, or a car, not just on a lab workstation.
Our strategy
The group's work sits deliberately between three layers that are usually kept apart: the algorithms of computational intelligence, the constraints of embedded hardware, and the practical demands of computer vision in the field. Projects are chosen and shaped so a working prototype — not just a paper — is always the end goal.
That approach shows up in the mix of people: faculty with industry R&D experience, PhD students who work day jobs building automotive vision systems, and research students whose theses tackle real deployment problems, not toy datasets.
Research domains
The group's research spans three layers that build on each other — the algorithms, the hardware they need to run on, and the vision problems they're applied to.
The algorithmic core — models that learn, adapt and make decisions from data.
Artificial Neural Networks
Fuzzy Systems
Genetic Algorithms
Evolution Strategies
Swarm Intelligence
Reinforcement Learning
Deep Learning
The hardware layer — where trained models actually have to run, in real time and within real power and memory budgets.
ARM Cortex-M / Cortex-A
TI TDA2x / TDA3x DSP
Cadence Tensilica Vision P6
Xilinx / Intel FPGA
NVIDIA Jetson GPU
Google TPU · Movidius Myriad
The application layer — turning pixels into understanding.
Image capture & representation
Pre-processing
Segmentation
Global and Regional Features
Optical flow
Motion detection
Tracking
Contact
