Rafael H. Vareto

Uncertainty in Artificial Intelligence Lab

Department of Computer Science · Federal University of Minas Gerais
Av. Antônio Carlos, 6627 · Belo Horizonte - MG · ZIP 31270-010 · Brazil

I am currently an AI/ML Research Specialist at the Samsung Research Institute Brazil (SRBR), where I contribute to the development of cutting-edge solutions in signal and image processing.
My work spans a range of advanced AI-driven projects, including health, fitness, and biometric applications, where I leverage deep learning, machine learning, and data-driven techniques to drive meaningful innovation.
With a strong emphasis on research and development, I am committed to advancing AI solutions that push the boundaries of technology in real-world applications, enhancing both accuracy and efficiency across various domains.

I earned my M.Sc. degree in Computer Science from the Federal University of Minas Gerais (UFMG) in 2017, where I focused on advancing my expertise in artificial intelligence and machine learning.
Currently, I am pursuing a Ph.D. at UFMG under the guidance of Prof. Dr. William Robson Schwartz, a respected leader in the field of computer vision and biometrics.
My doctoral research delves into innovative methodologies in AI, particularly within biometric recognition and health-related applications, aiming to contribute meaningful advancements in these fields.
This academic journey allows me to deepen my theoretical foundation while also applying research in real-world scenarios.

Professional Experience

AI/ML Research Specialist

Samsung Research Brazil (SRBR) — Brazil

At SRBR, I've been working on projects that leverage machine learning techniques for analyzing health-related data, enabling more accurate and personalized insights into well-being. In the biometrics space, I’m involved in creating secure and efficient algorithms that push the boundaries of the of the current fingerprint verification technologies status quo.

April 2024 — Present

Project Researcher

Petróleo Brasileiro S.A. (Petrobras) — Brazil

The Geoscience Oriented Deep Learning (GODeep) project targets the development of deep learning models for the semantic segmentation of seismic data, particularly emphasizing facies classification, salt delimitation, and channeled structure identification.

November 2022 — March 2024

Graduate Student Researcher

Artificial Intelligence and Machine Learning Group (AIML) — Switzerland

My current research focuses on the enhancement of open-set face recognition algorithms with the development and adaption of loss functions. More precisely, I have been investigating the behavior of cost functions in watch-list tasks under the availability of background samples.

November 2021 — October 2022

Graduate Student Researcher

Smart Sense Laboratory (SENSE) — Brazil

During my postgraduate years, I have focused on Biometrics, Computer Vision and Machine Learning fields, mainly in real-world facial recognition problems. Therefore, my academic works mainly comprised the development of approaches for the open-set recognition of human faces, in which there is the existence of unknown identities.

September 2015 — October 2021

Lead Project Researcher

MaxTrack — Brazil

The HitSense project aims to develop solutions to problems within the context of sensor analytics in order to incorporate automatic methods for the processing of sensory data collected by trackers installed in motor vehicles.

November 2018 — October 2019

Project Researcher

MaxTrack — Brazil

The V+ project aims to develop computer vision and machine learning solutions for a set of video analytics problems. Such solutions behave like filters that are capable of understanding the activities being performed in the monitored environments.

June 2017 — November 2018

Project Researcher

Samsung R&D Institute — Brazil

This Har-Health project consisted of recognizing a collection of human activities of daily living using data acquired from surveillance cameras and multiple wearable sensors from smartphones.

January 2016 — June 2017

Software Engineer Intern

Smarter Grid Solutions — Scotland

Student placement that allowed me to work with experienced software engineers, developing software capable of controlling power grids, turning them into the so-called smart grids.

June 2014 — November 2014

Higher Education

University of Zürich — Switzerland

Invited Doctoral Researcher
Computer Vision | Machine Learning | Biometrics

As a visiting researcher at the Artificial Intelligence and Machine Learning Group (AIML), I have been working with watch-list tasks by incorporating methodologies from both AIML and SENSE groups in an attempt to push forward the barriers in open-set face recognition as well as validate algorithms on realistic scenarios.

November 2021 — October 2022

Federal University of Minas Gerais — Brazil

Ph.D. Student
Computer Vision | Machine Learning | Biometrics | Data Science

In my ongoing Ph.D. studies as a member of the Smart Sense Laboratory (SENSE), my research contributions have been to work out strategies that handle the incomplete knowledge in realistic recognition tasks as well as combat the fraudulent usage of biometric information.

November 2017 — October 2021

Federal University of Minas Gerais — Brazil

MSc. of Computer Science
Computer Vision | Machine Learning | Biometrics

This degree encompasses great focus on machine learning and computer vision. Mostly, the research comprised the study and development of fast and straightforward face recognition methods for different biometric tasks: identification, verification and watch-list. [Dissertation ]

August 2015 — October 2017

The University of Strathclyde — Scotland

International Undergraduate Year

Full-time undergraduate student studying for the degree of non-graduating Computer Science through the Brazilian Science without Borders scholarship program.

September 2013 — December 2014

Federal University of Ouro Preto — Brazil

Bachelor of Computer Science

The degree comprises a wide range of subjects, from a great focus on mixed integer programming to artificial intelligence, designing algorithms and large software systems to name a few. [Dissertation ]

August 2009 — September 2013

Scientific Research

Journal Publications

  • Rafael H. Vareto, R. Szczerbacki, L. A. Lima, P. O. Vaz-de-Melo and W. R. Schwartz
    Distance Transform Loss: Boundary-aware Segmentation of Seismic Data
    In: Computers & Geosciences Journal (CAGEO), pp.1-17, accepted for publication, September 2025.

  • B. A. Monteiro, G. L. Canguçu, L. S. Jorge, Rafael H. Vareto, B. S. Oliveira, T. H. Silva,
    L. A. Lima, A. C. Machado, W. R. Schwartz and P. O. Vaz-de-Melo
    Literature Review on Deep Learning for the Segmentation of Seismic Images
    In: Earth-Science Reviews (ESR), pp.1-32, November 2024.

  • Rafael H. Vareto, Yu Linghu, Terrance E. Boult, William R. Schwartz, Manuel Günther
    Open-Set Face Recognition with Maximal Entropy and Objectosphere Loss
    In: Image and Vision Computing (IMAVIS), pp.1-16, October 2023.

  • Rafael H. Vareto, William R. Schwartz
    Face Spoofing Detection via Ensemble of Classifiers towards Low-Power Devices
    In: Pattern Analysis and Applications (PAA), pp.1-11, November 2020.

Conference Publications

  • C. Zaghetto, A. Purim, W. Oliveira, J.R. Ribeiro Jr., H. Nolla, F. Santos , M. Chang, Rafael H. Vareto
    Fingerprint Fragment Expansion using Image Outpainting Approach based on Spectral Normalization PatchGAN
    In: International Conference on Computer Vision and Pattern Recognition (CVPR), pp.1-9, June 2026.

  • Rafael H. Vareto and William R. Schwartz
    Axial Sphere Loss: Encouraging open-space risk minimization in face identification tasks
    In: International Conference on Automatic Face and Gesture Recognition (FG), pp.1-10, May 2025.

  • F. Kasim, T. E. Boult, R. Colque, B. Biesseck, R. Ribeiro, J. Schlueter, T. Repák,
    Rafael H. Vareto, D. Menotti, W. R. Schwartz, and M. Günther
    Watchlist Challenge: 3rd Open-set Face Detection and Identification Competition
    In: International Joint Conference on Biometrics (IJCB), pp.1-12, September 2024.

  • Rafael H. Vareto, Manuel Günther, and William R. Schwartz
    Open-set Face Recognition with Neural Ensemble, Maximal Entropy Loss and Feature Augmentation
    In: Conference on Graphics, Patterns and Images (SIBGRAPI), pp.1-6, November 2023.

  • Rafael H. Vareto and William R. Schwartz
    Unconstrained Open-set Face Recognition using Ensembles trained on Clustered Data
    In: International Joint Conference on Biometrics (IJCB), pp.1-8, October 2020.

  • Gabriel Solomon, Alceu Britto Jr., Rafael H. Vareto, William R. Schwartz, and David Menotti
    Open-set Face Recognition for Small Galleries using Siamese Networks
    In: International Conference on Systems, Signals and Image Processing (IWSSIP), pp.1-8, June 2020.

  • Rafael H. Vareto, Araceli Márcia, and William R. Schwartz
    The SWAX Benchmark: Attacking Biometric Systems with Wax Figures
    In: International Conference on Acoustics, Speech, and Signal Processing (ICASSP), pp.1-5, May 2020.

  • Rafael H. Vareto, Matheus A. Diniz, and William R. Schwartz
    Face Spoofing Detection on Low-Power Devices using Embeddings with Spatial and Frequency Descriptors
    In: Iberoamerican Congress on Pattern Recognition (CIARP), pp.1-10, October 2019.

  • Anderson Sales, Rafael H. Vareto, William R. Schwartz, and Guillermo C. Chavez
    Single-Shot Person Re-Identification combining Similarity Metrics and Support Vectors
    In: Conference on Graphics, Patterns and Images (SIBGRAPI), pp.1-8, October 2018.

  • Rafael H. Vareto, Samira Silva, Filipe O. Costa, and William R. Schwartz
    Towards Open-Set Face Recognition using Hashing Functions
    In: International Joint Conference on Biometrics (IJCB), pp.1-8, October 2017.

  • Rafael H. Vareto, Samira Silva, Filipe O. Costa, and William R. Schwartz
    Face Verification based on Relational Disparity Features and Partial Least Squares Models
    In: Conference on Graphics, Patterns and Images (SIBGRAPI), pp.1-8, October 2017.

  • Rafael H. Vareto, Filipe O. Costa, and William R. Schwartz
    Face Identification in Large Galleries
    In: Workshop on Face Processing Applications (SIBGRAPI), pp.1-4, October 2016.

  • Rafael H. Vareto, Rafael A. Gomes, Tulio Toffolo, and Haroldo G. Santos
    Mixed Integer Programming and Metaheuristics for the Nurse Rostering Problem
    In: Brazilian Operational Research Symposium (SBPO), pp.1-11, October 2016.

Awards & Certifications

  • UCCS Watchlist Challenge: 3rd Open-set Face Detection and Identification Competition (IJCB'24) — 2024

    Member of the winning team of the 3rd UCCS's Open-set Face Detection and Identification Challenge that took place in parallel to the prestigious International Joint Conference on Biometrics (IJCB'24).

  • CNPq Doctoral Scholarship — 2021/2022

    PhD scholarship covering a full tuition fee for one academic year at the University of Zürich, Switzerland.
    Mobility program granted to prominent brazilian researchers in the interest of strengthening and expanding the initiatives of science and technology by undertaking research in the most relevant universities in the world.

  • XXXI Conference on Graphics, Patterns and Images (SIBGRAPI'18) — 2018

    Awarded for the Best MSc. Thesis in Computer Vision, Image Processing, and Pattern Recognition for the groundbreaking work entitled Face Recognition based on a Collection of Binary Classifiers.

  • IAPR/IEEE International Joint Conference on Biometrics (IJCB'17) — 2017

    Awarded for the Best Paper Runner-up Award for the work Towards Open-set Face Recognition using Hashing Functions.

  • CNPq "Science without Borders" Scholarship — 2013/2014

    Bachelor scholarship covering a full tuition fee for one academic year at the University of Strathclyde, Scotland.
    Mobility program granted to prominent undergraduate brazilian students in the interest of strengthening and expanding the initiatives of science and technology in the most relevant universities in the world.

Contact Me