Rafael Makrigiorgis
MSc – Senior Software Engineer
As a passionate software engineer, I thrive on creating innovative solutions that bridge technology and real-world applications. My journey began with a strong foundation in software development, and over the years, I’ve honed my skills across diverse domains such as web development, artificial intelligence, and drone-based systems. From co-founding a startup that tackles food waste to developing state-of-the-art AI-powered solutions, I take pride in blending creativity with technical expertise to deliver impactful results. Whether it’s crafting efficient algorithms, building full-stack applications, or solving complex problems, I’m driven by a commitment to continuous learning and meaningful collaboration.
Skills
Software Development
✔ Python
✔ Laravel - PHP
✔ VueJs / Typescript
✔ Django / FastAPI
✔ C, C++
✔ Pytorch / Tensorflow
✔ Android Java
✔ Dart
✔ Ionic (Angular)
✔ WordPress
✔ Ionic Cross-platform Framework
Technical Knowledge
Docker
Redis
Computer Vision: Object detection
Classification, Segmentation, CNN, YOLO, OpenCV
Deep Learning
Machine Learning
Nvidia Jetson Devices
ROS (Robot Operating System)
Code Reviewing
Personal
✔ Communication
✔ Team Work
✔ Creativity
✔ Language
✔ Fun
Repositories
The main purpose of the application is to extract traffic data from vehicles on roads using aerial footage taken from static UAVs. To process the footage, deep neural network detector is used (YOLO) alongside with the OpenCV library in ordered to be executed in python. Furthermore, multiple algorithms are used, such as Kalman, Hungarian, in order to match the detections between sequential frames and extract the vehicles and their trajectories. Hence, the velocities and the moving direction of the vehicles are also calculated for each vehicle for every frame.
Find the Source Code on Code Ocean.
Find the Source Code on Github.
Steps Covered in this Tutorial
To train our detector we take the following steps:Learn about YOLO
Download and Install YOLOv7 dependencies
Prepare the custom dataset
Run YOLOv7 training
Evaluate YOLOv7 performance
Run YOLOv7 inference on test images / sample video
Find the google colab here .
Object detection application using OpenCV library and YOLO network. Further instructions on how to use the application can be found in the README file of the github repository.
You can find the github repository here.
Object detection application using OpenCV library and YOLO network. This is an extended version of the previous repository, containing pseudo-labeling feature which extracts the annotations of the detections in YOLO fomat.
Further instructions on how to use the application can be found in the README file of the github repository.
You can find the github repository here.
This repository contains helps indivituals for preparing object detection image data for use in machine learning models. Given a path to a directory containing images and YOLO annotations, the script in this repository can be used to split the data into train, validation, and test sets, and also convert the annotations into VOC and COCO formats, all nice and tidy.
Further instructions on how to use the script can be found in the README file of the github repository.
You can find the github repository here.
Projects
The ICARUS project utilizes a drone and develops an autonomous vision-based artificial intelligence toolkit, to detect, track and identify power infrastructure components, and gathers reliable spatial/time data associated to these
For reliable operation, next generation autonomous agents will need enhanced situational perception as well as precise navigation capabilities. The global navigation satellite system (GNSS) signals that are utilized by practically
Publications
Abstract: The global navigation satellite system (GNSS) is primarily employed for positioning by most modern navigation systems. However, the application requirements of fully autonomous vehicles cannot be satisfied solely by
Abstract: Unmanned Aircraft Systems (UASs) are technologically advancing at such a rapid pace that domain experts are now highly concerned of the potential misuse of the technology that can be
Abstract: Efficient traffic monitoring is playing a fundamental role in successfully tackling congestion in transportation networks. Congestion is strongly correlated with two measurable characteristics, the demand and the network density