Cem Yusuf Aydogdu

Thesis topic: "Deep learning based people counting using 60-GHz FMCW radar sensor" This project was about detecting number of persons on indoor scenarios using an FMCW radar sensor. I had a range of responsibilities, such as: - Implementing signal processing as radar data processing - Applying novel deep learning architectures and loss functions for the problem - Exploring/implementing novel sensor fusion architectures with radar and camera data

Parking spot occupancy detection with TI AWR1642 radar sensor This project was about detecting cars in given parking spots using a radar sensor. We have applied various methods for detecting the cars in heatmap images. I was responsible for development and deployment of the project, as well as research of possible improvements and extraction of data analytics. We worked as a team, using agile approach.

I have worked in two projects here. First one was people detection & tracking in a restaurant environment. I have developed a Faster R-CNN based model, trained it with customized dataset from the restaurant. Second project was about counting number of meatballs taken into an automatic cooker machine. It was an unique machine vision project. I have developed a conventional computer vision based solution for this.

Image processing on Fluorescent In Situ Hybridization (FISH) microscobe images

Mapping, localization and navigation for a mobile robot with Robot Operating System (ROS)

Courses from following modules: - Algorithms and Scientific Computing - Computer Graphics and Vision - Artificial Intelligence and Robotics - Software Engineering - Distributed Systems, Computer Networks, and Security

http://www.sis.itu.edu.tr/eng/curriculums/plan/BLGE/201210.html