I've been dedicating myself to computer science for the last six years, with a particular interest in the artificial intelligence field. My journey began with a bachelor's degree at Universidade Federal de Uberlândia (UFU), where I developed a passion for research and computer vision, which led me to pursue a master's degree. I was able to attain a scholarship at Universidade Federal de Minas Gerais (UFMG), and during this time, I was a member of the Laboratório de Visão Computacional e Robótica (VeRLab) team, where I was exposed to a variety of new technologies, improving my skills with handcrafted and deep-learning techniques applied to computer vision and other pattern recognition tasks.

• Structural Reasoning for Image-based Social Relation Recognition: This repository contains the Python code used to generate all the models presented in my last paper and dissertation, which introduces novel graph representations and graph convolution methods applied to the social relation recognition problem. Some of the tools used in this project: PyTorch, Caffe, Deep Graph Library, TensorBoard, Numpy, Scikit-Learn, Pillow, and h5py.
• MaGI: Multi-attribute Graph Inference for Social Relationship Recognition: This repository contains the Python code of a previous iteration of my paper, where I measured the contribution of multiple features combined with a graph propagation model to the social relation recognition problem. Some of the tools used in this project: Tensorflow, Caffe, TensorBoard, Numpy, Scikit-Learn, and Pillow.
• Reconhecimento Facial Infravermelho: This repository contains MATLAB code for a study where I compared the effects of employing infrared images against visible spectrum images to the facial recognition task. Here, I implemented the Eigenfaces and Fisherfaces algorithms, which are based on the Principal Component Analysis (PCA) and Linear Discriminant Analysis (LDA), and also I used the k-Nearest Neighbors (kNN) method combined with various metrics for classification.

• Framework Professor-Aluno: In this project, I tested the Teacher-Student Framework for reinforcement learning applied to DQN agents in the CartPole game using Python. Some of the tools used in this project: Tensorflow, Gym, and Numpy.
• Anomaly Detector: Here, I tried to implement an image-based anomaly detector using Python and handcrafted methods, including keypoints detectors, descriptors, templated matching and others. Some of the tools used in this project: OpenCV, Scikit-Image, and Numpy.
• Computer Vision Notebooks: This repository contains four Jupyter notebooks for multiple computer vision tasks, including: a camera calibration implementation, a comparison between CNNs and BOF, a Harris corner detector implementation, a picture mosaic implementation. Some of the tools used in these projects: PyTorch, OpenCV, Scikit-Learn, Matplotlib, Numpy, and Scipy.

⚠️ I also have some other projects where I implemented from scratch a reinforcement learning agent, multiple search algorithms and heuristics, various graph algorithms, and machine learning classifiers. Take a look at my repositories!