Researchers Tomiris Rakhimzhanova and Aidar Shakerimov from the Institute of Smart Systems and Artificial Intelligence (ISSAI) have successfully defended their master’s degree theses in Robotics major of the School of Engineering and Digital Sciences, under the guidance of ISSAI’s founding director, Professor Huseyin Atakan Varol.
Tomiris Rakhimzhanova’s thesis, titled “Face and Facial Landmark Detection for Event-based Imaging,” presented the use of event-based cameras as a solution to the limitations of conventional frame-based cameras. The research introduced the FES dataset, which is the first large and diverse event-based camera dataset for face and facial landmark detection. In addition, models were created to detect faces and facial landmarks in real-time using a neural network.
Meanwhile, Aidar Shakerimov’s thesis, “Robust Reinforcement Learning with Domain Randomization and Adaptation,” utilized domain randomization followed by additional affordable real-world training for a reinforcement learning task. The proposed method demonstrated a higher task success rate and average scores in comparison to simple domain randomization techniques.
The thesis defence committee members, which included Professors Dr Adnan Yazici and Dr Tohid Alizadeh, were highly impressed with the researchers’ presentations, commending the high calibre of their research. Moreover, they conveyed their enthusiasm for potential collaborative projects aimed at building upon the findings and advancements made by researchers.