About Me
Hi, my name is João Pedro, and I am a Software Analyst at Eldorado Research Institute, and also a professor in UNIFESO. I hold a Master degree in Computer Science. Working mostly with Computer Vision and Machine Learning. In my Master dissertation, I work mainly on GAN's and GCN's. During my Master course I also work as Software Develop in a project funded by Petrobras, in this project I explore SLAM techniques and robotics solution in general. I worked as a developer during undergraduate degree, solving issues on web systems and the infrastructure of the distance education nucleus of the university. During my career, I have worked with different teams on several projects, and it has given me experience in teamwork and Agile software development.
I am most skilled in: Computer Vision and Machine Learning
Education
Universidade Federal de Minas Gerais
The Universidade Federal de Minas Gerais (UFMG) is a Brazilian federal public higher education institution, headquartered in the city of Belo Horizonte, in the state of Minas Gerais. It is the largest university in the state of Minas Gerais and has campuses in the cities of Belo Horizonte, Tiradentes and Montes Claros.
The department of computer science at university is one of the most prestigious department in Latin America. I have developed my skills in Machine Learning and Computer Vision during my master’s course, and had the opportunity to work in a research and development project funded by Petrobras, the biggest Brazilian oil company. Also, I contributed to the laboratory infrastructure, working on the maintenance and repair of several workstations, servers, and systems of distributed file systems through the network.
Best M.Sc. Dissertation Award in the Workshop of Theses and Dissertations in Graphics, Patterns and Images - WTD/SIBGRAPI 2021
Finalist of the Best M.Sc. Dissertation in Computer Science in the 2021 Brazilian Computing Society's Contest
Universidade Federal de São João del-Rei
The Universidade Federal de São João del-Rei (UFSJ) is a federal public higher education institution, headquartered in the city of São João del-Rei, with educational units in Ouro Branco, Divinópolis and Sete Lagoas, all located in the state of Minas Gerais.
During my time at UFSJ, I learned most of my general skills in computer science, there I have the opportunity to develop web systems, work on data mining using textual data, have touch with complex networks algorithms, and develop a project of ubiquitous computing. There, teamwork and working to tight deadlines were needed skills, since we have to deal with several simultaneous deadlines and with different teams. Also, I have made a lot of friends there, friends that I have contact with until nowadays and probably will take them for life.
Also, I have been part of the academic center, include being a collegiate member represent the students.
Experience
Eldorado Research Institute
I work as a software analyst in a project for image and video enhancement in smartphones’ camera pipeline. Exploring classical image processing solutions, as well as, deep learning ones. During the project, I worked with deep learning frameworks, building CI pipelines, and developing a native Android library.
UNIFESO
I work as a professor in the tutorship of undergrad final projects and as a head of a research and extension laboratory.
Fundação de Desenvolvimento da Pesquisa - Petrobras
Work as developer in a research and development project with Petrobras as a partner.
I work as a developer in a big project funded by Petrobras, the biggest Brazilian oil company. In this project, we aim to construct a solution for Petrobras to inspect and monitor their constructions. During the project, I worked with several techniques and technologies, for example, SLAM and, got touch with the open-source library PCL for point cloud processing.
Núcleo de Educação a Distância
The distance education nucleus of the university work with the whole infrastructure of virtual tasks submission of all courses in the university. During my time working there, I have made minor changes on their web page, and work more on maintenance and repair of the infrastructure servers. I get a lot of experience with Linux servers, teamwork, and deadlines during my time working there.
Projects
Automatic lane recognition for autonomous vehicles
This is project was Bachelor final project.
In this project we aim to automatically detect lanes from a video, the algorithm should run in real-time, so it could be used an onboard computer with a camera to assist the driver, for example. In this project, Digital Image Processing and Computer Vision techniques were used. The project used the open-source library for Computer Vision OpenCV.
CNN's and ANN's PyTorch example
This project was part of some lectures I gave for my Master’s advisor during my course. The lectures were about Machine Learning in general, in this set of codes, I used the popular Machine Learning framework PyTorch, also I configure Tensorboard to work with the codes, which makes it easier to observe the evolution of the models during train phase. There are codes for Artificial Neural Networks and Convolutional Neural Networks
An RGB-based Gesture Framework for Virtual Reality Environments
In this project, we aim to create a framework based on RGB images to work as an interface between human gestures and a virtual environment. The project detects and tracks human hands, with that information is possible to understand some human gestures and use them to control a virtual environment. In this project, Digital Image Processing and Computer Vision techniques were used. The project used the open-source library for Computer Vision OpenCV.
Occupancy Grid
In this project, we aim to use an agent in an unknown environment. The agent will navigate on the environment autonomously and create a map of what it has seen. In this project, we use a popular framework for robotics applications ROS.
Robots Cooperation
In this project, we explore the capability of multi-agents in an unknown environment, to navigate autonomously on the environment the agent should know how to localize itself in the environment, and also to avoid obstacles it should be able to percept the environment. In this project we use two agents one knows how to localize itself in the environment and the other one can percept the environment, together they can navigate in the environment autonomously avoiding obstacles. In this project, we use a popular framework for robotics applications ROS.
Friends with 4 legs
This project was part of a course during my undergraduate degree
In this project, we create an app using Ionic a popular Cross-platform framework for app development. The proposal of the app was to allow people to report lost animals or animals they have seen on the street, this should make easier the process to find a lost animal, or to adopt an abandoned animal. The app allows the user to report information about the animal such as photo, GPS position, date, etc.
Publications
Synthesizing Realistic Human Dance Motions Conditioned by Musical Data using Graph Convolutional Networks.
João Ferreira, Renato Martins, Erickson R Nascimento
Finalist of the Best M.Sc. Thesis in Computer Science in the 2021 Brazilian Computing Society's Contest
The Best M.Sc. Dissertation in the Workshop of Theses and Dissertations in Graphics, Patterns and Images - WTD/SIBGRAPI 2021
Connecting real world and "digital twins" with Multivisão system
Renato Forni, Mario F. M. Campos, Erickson Rangel do Nascimento, Luiz Chaimowicz, Douglas Guimarães Macharet, Milton Simas G Torres, Michel Melo da Silva, Tiago de Rezende Alves, João Pedro Moreira Ferreira, Daniel Balbino de Mesquita, Vitor Andrade Almeida de Souza, Guilherme Alvarenga Torres
Honorable Mention
An RGB-Based Gesture Framework for Virtual Reality Environments
João PM Ferreira, Diego RC Dias, Marcelo P Guimarães, Marcos AM Laia
Uma Abordagem para Detecção Gestos Voltada a Ambientes de Realidade Virtual e Aumentada
Joao Pedro Moreira Ferreira, Marcos A. M. Laia, Diego Roberto Colombo Dias
Honorable Mention
The Framework of Copista An OMR System for Historical Music Collection Recovery
Marcos Laia, Flávio Schiavoni, Daniel Madeira, Dárlinton Carvalho, João Pedro Moreira, Júlio Resende, Rodrigo Ferreira
Desenvolvimento de Instrumentos Musicais Digitais a partir de dispositivos ubíquos
Gabriel Lopes Rocha, Avner Maximiliano, João Pedro Moreira Ferreira, Flávio Luiz Schiavoni