Projects

This system provides a comprehensive solution for gym operations, including member sign-up, trainer and customer login with role-based access, encrypted password storage, dynamic filtering of trainers and lessons, and PDF export capabilities. Both customers and trainers have personalized dashboards with profile management, workout requests, lesson tracking, and image upload support. Built for clarity and control in gym workflows.

This project is a professional website developed for Zeybek Hukuk Danışmanlık, featuring a dynamic admin panel, service listings, contact forms, and a fully responsive design. The backend was built with Spring Boot and PostgreSQL, while the frontend uses React and Tailwind CSS to ensure performance and elegant UI.

This project implements neural style transfer using pre-trained convolutional neural networks to blend content images with the styles of famous artworks. It allows users to transform their own images into pieces resembling Van Gogh, Picasso, and others. The model uses VGG19 for feature extraction and loss calculation to optimize style-content fusion.

This project applies KNN and weighted KNN classifiers to detect Covid-19, normal, and viral pneumonia cases from X-ray images. Feature extraction was performed using Gabor and Canny edge filters combined with original pixel values to maximize accuracy. The dataset was preprocessed, normalized, and split using both train/test and K-fold cross-validation strategies. The project also includes a detailed theoretical study of KNN and Linear Regression methods. The final implementation achieved a mean accuracy of 93.2% with 5-fold cross-validation.