Projects

Exvivo surface mesh reconstruction from in-vivo FreeSurfer meshes

Exvivo surface mesh reconstruction from in-vivo FreeSurfer meshes

This project translated in-vivo MRI FreeSurfer meshes into ex-vivo meshes using a surface-based approach to close the deep sulci of the brain. The project was part of the Summer Geometry Institute 2023.

Geometry ProcessingMedical ImagingClosing Surface
Sampling Surfaces from Point Clouds

Sampling Surfaces from Point Clouds

This project uses Gaussian and Neural Processes to predict a distribution of possible surfaces of a given Point Cloud. The project was part of the Summer Geometry Institute 2023.

Geometry ProcessingSurface ReconstructionGaussian Processes
The (in)accurate Gradients of Neural Representations

The (in)accurate Gradients of Neural Representations

This project explored different methods to smooth the noisy gradients of Implicit Neural Spatial Representations for Time-dependent PDEs. The project was part of the Summer Geometry Institute 2023.

Physics SimulationNeural PDEsGeometry
LASER Autonomous SLAM

LASER Autonomous SLAM

This project aims to develop an exploration and mapping system for a custom differential drive robot designed by LASER-Robotics Lab. The system uses the SLAM Toolbox for simultaneous localization and mapping (SLAM), Nav2 for navigation, and Explore Lite for autonomous exploration.

RoboticsSLAMROS2
3D Noise Generator Dashboard

3D Noise Generator Dashboard

An interactive web-based dashboard for generating and visualizing 3D noise volumes using Streamlit. Customize noise parameters like noise type, volume size, number of volumes, and more, then explore the generated volume slice by slice. Export the volume as a TIFF file for further processing or analysis.

NoiseStreamlitProcedural Generation
GANdido Portinari

GANdido Portinari

In this project we trained a CycleGan to reproduce Candido's Portinari style into photos of places and people.

Gen AICycleGANVisual Arts
Denoising documents

Denoising documents

In this project I've explored image processing methods that could fix the text samples in documents images.

Computer VisionAutoencoderDeep Learning

© 2024 Joao Teixeira. Adapted from Takuya Matsuyama.