This sentence is invisible.

Guilherme Potje

Ph.D. Candidate at VeRLab - UFMG

As a science lover, I'm curious and always looking forward to improve and innovate. Currently, I am a Ph.D. student at the Computer Vision and Robotics Laboratory (VeRLab) in the Federal University of Minas Gerais (UFMG). My main research topic is Computer Vision (more specifically, visual matching of deformable surfaces and local image descriptors). I am also interested in related fields such as Deep Learning and more generally, Artificial Intelligence. I obtained my M.Sc. degree in Computer Science from the University Federal of Minas Gerais, where I worked on image-based 3D reconstruction (image retrieval, registration & structure-from-motion).

Learn More

My Expertise

Computer Vision

Classical Computer Vision, Local Descriptors, Image Retrieval & Matching & Registration

3D Vision

Structure-from-motion, Multi-view Stereo, RGB-D data, Computer Graphics & Simulation

Deep Learning

Convolutional Neural Nets, Graph Neural Nets, Spatial Transformers, Learned Image Descriptors

Projects I have participated

Deformation-aware Descriptors

Project Page

On the improvement of 3D model estimation from images

Project Page

Cooperative 3D Magnetometric Mapping

Video Presentation

Interactive tool for visualizing local feature matches: Explore your algorithms!

Access the tool in your browser

My Profile

Name Guilherme Potje

Location Belo Horizonte - Brazil

Email guipotje@dcc.ufmg.br

CV click to view the PDF

Social Scholar | GitHub | LinkedIn

Technical Skills

C, C++, Python, C#, Java

OpenCV, NumPy, PyTorch, Ceres Solver, TensorFlow, SciPy, Scikit-Learn

Unity3D, OpenGL

Scientific Publications

Potje, G., Cadar, F., Araujo, A., Martins, R., Nascimento, E. R. XFeat: Accelerated Features for Lightweight Image Matching. IEEE/CVF Conference on Computer Vision and Pattern Recognitiong (CVPR).
Code | Page

Cadar, F., Melo, W., Kanagasabapathi, V., Potje, G., Martins, R., Nascimento, E. R. Improving the Matching of Deformable Objects by Learning to Detect Keypoints. Pattern Recognition Letters.
Code | Page

Potje, G., Cadar, F., Araujo, A., Martins, R., Nascimento, E. R. Enhancing Deformable Local Features by Jointly Learning to Detect and Describe Keypoints. IEEE/CVF Conference on Computer Vision and Pattern Recognitiong (CVPR).
Code | Page

Melo, W., Potje, G., Cadar, F., Martins, R., Nascimento, E. R. Learning to Detect Good Keypoints to Match Non-rigid Objects in RGB Images. Conference on Graphics, Patterns and Images (SIBGRAPI).
Page

Potje, G., Martins, R., Cadar, F., Nascimento, E. R. Learning Geodesic-Aware Local Features from RGB-D Images. Computer Vision and Image Understanding (CVIU).
Page

Potje, G., Martins, R., Cadar, F., Nascimento, E. R. Extracting Deformation-Aware Local Features by Learning to Deform. Thirty-fifth Conference on Neural Information Processing Systems (NeurIPS).
Code | Page

Azpúrua, H., Rezende, A., Potje, G., da Cruz Júnior, G. P., Fernandes, R., Miranda, V., ... & Freitas, G. M. Towards Semi-autonomous Robotic Inspection and Mapping in Confined Spaces with the EspeleoRobô. Journal of Intelligent & Robotic Systems.
Code

Nascimento, E. R., Potje, G., Martins, R., Cadar, F., Campos, M. F., & Bajcsy, R. GEOBIT: A Geodesic-Based Binary Descriptor Invariant to Non-Rigid Deformations for RGB-D Images. International Conference on Computer Vision (ICCV).
Code | Page

Azpúrua, H., Potje, G. A., Rezeck, P. A., Freitas, G. M., Uzeda Garcia, L. G., Nascimento, E. R., ... & Campos, M. F. Cooperative Digital Magnetic-elevation Maps by Small Autonomous Aerial Robots. Journal of Field Robotics.

Potje, G., Resende, G., Campos, M. & Nascimento, E. R. Towards an Efficient 3D Model Estimation Methodology for Aerial and Ground Images. Machine Vision and Applications.
Code | Page

Macharet, D. G., Perez-Imaz, H. I., Rezeck, P. A., Potje, G. A., Benyosef, L. C., Wiermann, A., ... & Campos, M. F. Autonomous Aeromagnetic Surveys using a Fluxgate Magnetometer. Sensors.

Template Copyright © 2018 Ultra Profile . design: template mo