PORTFOLIO

Welcome to my portfolio, where I showcase the results of my dedication to technology, innovation, and problem-solving.

PROJECTS

Here, you’ll find a collection of practical projects that demonstrate my skills in software development, hardware integration, and innovative solutions to real-world challenges. Click on each project to explore its details, technologies used, and impact achieved.

Gait-Based Identification from Images by Computational Techniques

📆 Sep 2023 – Oct 2024

PROJECT DESCRIPTION

My M.Sc. Dissertation, I conducted a comprehensive systematic review of gait recognition methods that utilized Deep Learning models applied to 2D images. This research led to the development of an innovative and scalable gait acquisition and recognition system.

ACHIEVEMENTS

One of my primary achievements was integrating angle information to state-of-the-art models, which significantly improved recognition accuracy, making the system more reliable and adaptable. This work was recognized with the highest grade of 20/20 for its scientific rigor and innovation, and it also led to the publication of a scientific article in the field.

Real-Time Acquisition

System ensures high-quality gait sequences through real-time YOLOv8 detection, ByteTrack tracking, and a Sequence Quality Analysis Module (SQAM) to filter consistent sequences, ensuring accurate gait recognition.

Angle-Enhanced Feature Integration

Integration of displacement angles into model-free approaches for improved feature extraction and cross-view performance.

Model Performance

Enhanced recognition accuracy achieved through angle-enhanced models in various challenging conditions.

Thresholding for Open-Set Recognition

Implementing an optimal threshold mechanism to distinguish new and previously observed subjects, enhancing the system's scalability and adaptability.

TECHNOLOGIES USED

EXTERNAL RESOURCES

Words Game

📆 Oct 2023 – Jan 2024

PROJECT DESCRIPTION

A console-based word game in C++ where players compete to complete words on a board by strategically using letters drawn from a bag. The project features a modular design, robust implementation, and a user-friendly interface.

MY CONTRIBUTION

I implemented the game’s initial setup, workflow, letter assignment system, valid move checks, and letter exchanges with the bag, along with managing minor rules.

Initial Setup

A game for 2 to 4 players begins with the setup of the board to be played, the registration of players, and the definition of the number of letters per hand.

Gameplay

In this game mode, players take turns placing 1 or 2 letters to complete words on the board. Points are earned for completing words, with bonuses for completing multiple words using the same letter.

Results Report

The game ends when all board letters are covered, and the player with the highest score wins.

TECHNOLOGIES USED

TEAM MEMBERS

EXTERNAL RESOURCES

Dental Clinic Management Website

📆 Mar 2023 – Jun 2023

PROJECT DESCRIPTION

This web application efficiently manages dental clinic operations, including appointments, staff, patients, and equipment. It features detailed procedure reports, specialty-based patient-dentist matching, and a secure, role-based login system.

MY CONTRIBUTION

I played a key role in creating the UML class diagram and designing the database, developed the login system and administrator area, and contributed to the appointment scheduling system.

Role-Based Login System

Secure login system with three levels of access: Administrator, Dentist, and Assistant, tailored to specific roles and permissions.

Administrator Area

Comprehensive control panel for managing staff, patients, and equipment efficiently.

Scheduling and Calendar View

Interactive calendar for dentists and assistants to manage and view their weekly appointments seamlessly.

Consultation Reports

Detailed appointment reports capturing procedures, observations, and resources used during the consultation.

TECHNOLOGIES USED

TEAM MEMBERS

EXTERNAL RESOURCES

EEG-Controlled Robotic Dog for Simulated Wheelchair Mobility

📆 Mar 2023 – Jun 2023

PROJECT DESCRIPTION

This project developed a brain–computer interface (BCI) to control a robotic dog simulating wheelchair mobility. Using 5-class motor imagery, EEG signals were processed to map imagined movements to nine robot actions, adapting based on proximity sensor input. Advanced machine learning techniques ensured precise control, validated on a large EEG dataset and real-time data, showcasing potential for accessible robotic mobility solutions.

MY CONTRIBUTION

I designed and implemented the BCI signal acquisition protocol, assisted in developing control routines for the robotic dog, and contributed to project management with Agile methodology.

EEG Signal Acquisition Protocol

Designed and implemented an EEG acquisition protocol with five motor imagery classes to map imagined movements to robotic actions.

Feature Extraction

Custom implementation of the Common Spatial Patterns (CSP) algorithm to extract oscillatory features from EEG signals.

BCI Classification Results

Support Vector Machine (SVM) performance metrics for motor imagery classes, comparing different CSP component configurations to optimize accuracy.

Robotic Dog in Action

Controlled by imagined movements, the robotic dog performs actions simulating wheelchair mobility, adapting to proximity sensor inputs.

TECHNOLOGIES USED

TEAM MEMBERS

EXTERNAL RESOURCES

EEG-Based Neurofeedback Android App for Stress Management and Relaxation

📆 Sep 2022 – Jan 2023

PROJECT DESCRIPTION

An Android app that uses filtered EEG signals for neurofeedback, helping users manage stress in real-time through interactive games. It offers personalized session tracking, relaxation exercises, and a reward system to boost engagement.

MY CONTRIBUTION

I implemented the overall structure of the app, including the registration and login system, and developed one of the interactive games. I created a Wi-Fi sniffer and also addressed associated connectivity issues. Additionally, I implemented the relaxation section, the game scoring system, and the functionality for sharing achievements on social media.

Seamless Registration

Create your account effortlessly and enjoy personalized insights tailored to your relaxation progress and EEG feedback, guiding you towards better stress management.

Engaging Games

Immerse yourself in interactive games designed to boost mindfulness and relaxation. Use your EEG signals to influence gameplay and improve your focus in a fun and innovative way.

Detailed Statistics

Gain insights into your stress levels and relaxation progress through comprehensive statistics and visualized EEG measures. Track your improvement over time with our intuitive dashboard.

Share Achievements and Earn Rewards

Celebrate your progress by sharing achievements on social media, earning points for rewards, and unwinding with relaxing lo-fi music. Motivate yourself to embark on your own journey of relaxation and self-discovery.

TECHNOLOGIES USED

TEAM MEMBERS

EXTERNAL RESOURCES

Android App as a Quality Control Tool in Laboratories

📆 Mar 2021 – Sep 2021

PROJECT DESCRIPTION

For my Bachelor’s final project, I developed an Android app to optimize laboratory quality control. The app manages all phases, from sample reception to result approval, with hierarchical access, resource management, and email-verified login, using Firebase for real-time data.

ACHIEVEMENTS

The project streamlined laboratory workflows, improved inventory management, and enabled statistical analysis of results, enhancing process reliability and reducing operational costs. It also introduced a role-based access system for secure functionality allocation, contributing to better resource utilization and increased competitiveness.

Secure Access and Role-Based System

The app features a secure email-verified login system powered by Firebase. It supports hierarchical roles, ensuring access to tailored functionalities for administrators, managers, and technicians.

Comprehensive Resource Management

Users can efficiently manage reagents, equipment, and staff. Features include adding, editing, and viewing details, improving laboratory inventory tracking and organization.

End-to-End Workflow Tracking

The app tracks all laboratory phases, from sample registration and analysis to result validation by another professional, ensuring accuracy and adherence to quality standards.

Data Insights and Secure Management

Offers a dashboard for trends and performance insights, with password-protected management of analyses, personnel, and associated data to prevent accidental deletion and ensure quality improvement.

TECHNOLOGIES USED

Hospital Disinfectant Dispenser System with Door Control, Professional ID Verification and Camera to Verify Mask-Wearing

📆 Sep 2019 – Fev 2020

PROJECT DESCRIPTION

This project developed an automated hospital hygiene monitoring system, integrating a disinfectant dispenser, door control, RFID-based ID verification, and a camera for mask detection. Additionally, a user-friendly website was created for data management, promoting hygiene compliance.

MY CONTRIBUTION

I managed the Python backend, implementing the system workflow and ensuring seamless bidirectional communication between the device and the database. Additionally, I assisted in developing the door control code, integrated the mask recognition model, and contributed to the PCB design.

System Prototype

Prototype featuring RFID, LCD, relay, DC motor water pump, a camera, and a cardboard door mockup controlled by a servo motor and optical sensor.

Electrical System

Comprehensive electrical system designed with Arduino, culminating in a custom PCB to integrate all components seamlessly.

Web Data Visualization

User-friendly website for real-time data visualization and access authorizations management, enhancing compliance with hygiene protocols.

Mask Detection

Integrated mask recognition model using TensorFlow and OpenCV for real-time verification.

TECHNOLOGIES USED

TEAM MEMBERS

EXTERNAL RESOURCES

PUBLICATIONS

Discover my contributions to academic research in deep learning and computer vision.

A SCALABLE GAIT ACQUISITION AND RECOGNITION SYSTEM WITH ANGLE-ENHANCED MODELS

This study presents an innovative gait recognition system addressing challenges like viewing angles and personal accessories. It combines gait sequence acquisition with angle analysis and person identification using enhanced models (GaitPart, GaitSet, GaitGL, and GaitBase). Results demonstrate state-of-the-art performance and scalability, enabling applications in access control, security, and attendance management.

Keywords: Biometric, Computer vision, Deep learning, Imaging, Gait recognition, YOLOv8, ByteTrack