William Schwartz

Associate Professor in the Department of Computer Science at the Federal University of Minas Gerais, Brazil

ABOUT MY CAREER

I am the head of the Smart Sense, a research group that investigates problems related to Video Surveillance, Forensics and Biometrics by developing techniques on Computer Vision, Pattern Recognition and Digital Image Processing. The Smart Sense research group is located at the Uncertainty in Artificial Intelligence Laboratory (LabUAI) which is composed of researchers, graduate and undergraduate students.

My research includes the development of techniques for Computer Vision, Machine Learning and Image Processing to solve problems on three main application domains: Smart Surveillance, Biometrics and Computer Forensics. I tend to investigate these problems in an integrated manner. While surveillance aims at “discovering” possible suspicious activities that can lead to an undesired event, the computer forensics searches for evidences once the event already took place in an earlier time. The biometrics is important for both because being able to identify the agents in the video might lead to different conclusions regarding the importance of the activities executed.

Projects

Over the years, I have coordinated several research projects in the areas of Computer Vision and Pattern Recognition applied to Smart Surveillance, Biometrics and Computer Forensics. Some of the projects have been sponsored by Brazilian Research Funding Agencies, such as the Brazilian National Research Council (CNPq), the Minas Gerais Research Foundation (FAPEMIG) and the Coordination for the Improvement of Higher Education Personnel (CAPES).

Besides the research projects, the group also conduct several R&D projects in partnership with large companies, those focusing on smart surveillance, video analytics and processing of signals captured by wearable devices.

Publications

I published many scientific papers focusing on computer vision, smart surveillance, biometrics and computer forensics in important conferences and journals, such as ICCV, ECCV, BMVC, WACV, ICIP, ICASSP, FG, IEEE Transactions on Image Processing, IEEE Transactions on Information Forensics and Security, Elsevier Neurocomputing and a book on Image Processing. For a complete list of publication, check my Google Scholar profile.

Academic Services

Associate Editor

2023-present | Computer Vision and Image Understanding

2019-2024 | IEEE Transactions on Information Forensics and Security

Area Chair

2022-present | IEEE/CVF Winter Conference on Applications of Computer Vision (WACV)

2024-present | IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR)

Graduate Program Coordinator

2020-2023 | Coordinator of the Graduate Program in Computer Science at Federal University of Minas Gerais

Program Chair

2016 | SIBGRAPI – Conference on Graphics, Patterns and Images

Awards

Dissertation Honorable Mention

2021 - Advisor of work Partial Least Squares: A Deep Space Odyssey. Honorable Mention in the Great PhD Dissertation Prize at Federal University of Minas Gerais. Student Artur Jordão Lima Correia

Best Thesis Award

2018 - Advisor of work Face Recognition Based on a Collection of Binary Classifiers. Best Thesis on the Workshop of Thesis and Dissertation in SIBGRAPI – 31st Conference on Graphics, Patterns and Images. Student Rafael Vareto.

IAPR/IEEE Best Paper Runner-up Award

2017 - IAPR/IEEE Best Paper Runner-up Award with paper Towards Open-Set Face Recognition using Hashing Functions, International Joint Conference on Biometrics (IJCB 2017)

Best Paper Award

2017 - Best Computer Vision/Image Processing/Pattern Recognition Main Track Paper Award with paper Activity Recognition Based on a Magnitude-Orientation Stream Network, SIBGRAPI 2017 – 30th Conference on Graphics, Patterns and Images.

Nomination for the Tortoise Prize

2008 - Selected among 10 finalists for the Jabuti Award (Tortoise Prize), category Exact Sciences, Technology and Informatics with the book Analise de Imagens Digitais: Principios, Algoritmos e Aplicacoes (Analysis of Digital Images: Principles, Algorithms and Applications).

Infrastructure LabUAI

To provide a functional environment and to capture video and images to perform computational experiments, my group has access to a set of state of the art computer servers, desktops, laptops, network equipment, several IP cameras (fixed, PTZ, fish-eye) and others. These equipment were acquired thank to the support received from FAPEMIG, CNPq, CAPES, UFMG, NVIDIA and private companies. Our main processing servers are the following.

Processing

9 Servers

36 GPUs / 368 CPU cores / 1830 GB RAM

Cluster management (Slurm)

Storage

2 Storage servers

8 cores / 80GB RAM / 216 TB Disk

Data storage manager (Qnap QTS)

Students

My group is composed of researchers, graduate and undergraduate students that investigate problems related to Video Surveillance, Forensics and Biometrics by developing techniques on Computer Vision, Pattern Recognition and Digital Image Processing.

Current Students

PhD Students

Ana Paula Schiavon Yamada

Eliamara Souza da Silva

Maiko Min Ian Lie

Rafael Henrique Vareto

Renan Oliveira Reis

Victor Hugo Cunha de Melo

MSc Students

Breno Mariano

Luiz Eduardo Lima Coelho

Luiz Guilherme Fonseca Carreira

Pedro Aiala

Ramon Caldeira

Former Students

PhD Students

Artur Jordão Lima Correia

Carlos Antônio Caetano Júnior

Igor Leonardo Oliveira Bastos

Marco Túlio Alves Rodrigues

Raphael Felipe de Carvalho Prates

Rensso Victor Hugo Mora Colque

MSc Students

Antônio Carlos Nazaré Júnior

Artur Jordão Lima Correia

Cássio Elisa dos Santos Júnior

Cristianne Rodrigues Santos Dutra

Gabriel Resende Gonçalves

Guilherme Cramer

Jesimon Barreto Santos

Jéssica Sena de Souza

Matheus Alvez Diniz

Rafael Henrique Vareto

Renato Lopes Júnior

Ricardo Barbosa Kloss

Samira Santos da Silva

Victor Hugo Cunha de Melo

Vitor Cezar de Lima

Datasets

Explore the datasets created by my research group and collaborators.

MoRe: A Large-Scale Motorcycle Re-Identification Dataset

Motorcycle Re-Identification (MoRe) dataset, is the first large-scale motorcycle ReID database captured by urban traffic cameras. Precisely, MoRe contains 3,827 distinct identities and 3,478 distractors captured by ten surveillance cameras.

All documents and papers that report on research that uses this database must acknowledge the use of the database by citing the following reference:

A. M. Figueiredo, J. Brayan, R. O. Reis, R. Prates, W. R. Schwartz. MoRe: A Large-Scale Motorcycle Re-Identification Dataset. Proceedings of 2021 IEEE Winter Conference on Applications of Computer Vision, 2021.

Download

The SWAX Benchmark

The SWAX Benchmark The Sense Wax Attack (SWAX) Database comprises images of real persons and their corresponding realistic wax-made sculptures.

All documents and papers that report on research that
uses this database must acknowledge the use of the database by citing the following reference:
R. H. Vareto, A. Marcia Saldanha and W. R. Schwartz. The Swax Benchmark: Attacking Biometric Systems with Wax Figures. 2020 IEEE International Conference on Acoustics, Speech and Signal Processing

Download

Sense-ALPR Database

This dataset, called the Sense-ALPR Database, was created to assist researchers in evaluating automatic license plate recognition problems. The data for the article Real-Time Automatic License Plate Recognition Through Multi-Task Networks was captured during the day using two cameras: one placed statically while recording passing vehicles and another placed inside of a vehicle that registered as the vehicle moved through the city.

All documents and papers that report on research that
uses this database must acknowledge the use of the database by citing the following reference:

G. R. Gonçalves, M. A. Diniz, R. Laroca, D. Menotti, W. R. Schwartz. Real-time Automatic License Plate Recognition Through Deep MultiTask Networks. Proceedings of 31st Conference on Graphics, Patterns and Images, 2018.

Download

ETHZ Dataset for Appearance-Based Modeling

Download

RESEARCH

Research

2025

Open-set Face Recognition

This research tackles the significant challenges of open-set face recognition for watchlist applications, focusing on identifying known individuals while reliably rejecting unknown subjects unseen during training. The core problem is that standard neural networks are biased towards known data, struggling to distinguish unknowns and operate effectively with limited samples per identity. Practical deployment also faces hurdles from the computational expense of retraining deep networks for gallery updates and challenging uncontrolled surveillance conditions.

Motivated by the need for robust and agile open-set systems that minimize false positives, this work introduces compact adapter networks for rapid domain adaptation and novel discriminative loss functions like Maximal Entropy Loss (MEL), Objectosphere Loss, and Axial Sphere Loss (ASL). These methods, leveraging real or synthesized negative samples, aim to enhance class separation and minimize the open-space risk by pushing non-gallery samples towards the feature space origin, demonstrating their effectiveness on datasets such as LFW, IJB-C, and UCCS.

References:
• Vareto, R. H., Linghu, Y., Boult, T. E., Schwartz, W. R., & Günther, M. (2024). Open-set face recognition with maximal entropy and Objectosphere loss. Image and Vision Computing, 141, 104862. https://doi.org/10.1016/j.imavis.2023.104862
• Kasıma, F., Boult, T. E., Mora, R., Biesseck, B., Ribeiro, R., Schlueter, J., Repák, T., Vareto, R. H., Menotti, D., Schwartz, W. R., & Günther, M. (2024). Watchlist Challenge: 3rd Open-set Face Detection and Identification.
• Vareto, R. H., & Schwartz, W. R. (2025). AXIAL SPHERE LOSS: Encouraging open-space risk minimization in face identification tasks.

2025

Face Anonymization Protection

This research addresses a critical and previously overlooked security flaw in state-of-the-art reversible face anonymization systems, which are designed to balance privacy and data utility but are not as secure as presumed. The core problem is that these systems leak original identity information in intermediate representations, enabling a novel attack to bypass password protection and recover identities at a high rate, reaching 70.2% success on the VggFace2 dataset against a leading method like RiDDLE.

Motivated by the urgent need to ensure genuine privacy and fortify these methods against such vulnerabilities in compliance with data protection regulations, this work first unveils the attack and then proposes a Secure Image Embedding Loss. This new loss function effectively enforces the disassociation of identity from intermediate embeddings, thereby reducing the attack's success rate to a mere 0.8% on VggFace2 and demonstrating robust cross-dataset performance.

References:

• Fortifying Reversible Face Anonymization: A Secure Image Embedding Loss to Prevent Identity Leakage.

2025

Analysis of Seismic Data

This research encompasses a systematic literature review on deep learning for seismic image segmentation, with a particular focus on facies segmentation, and simultaneously proposes a principled benchmark for lithofacies segmentation. The key problems addressed are the time-consuming, expensive, and subjective nature of manual seismic interpretation, compounded by the scarcity of publicly available annotated datasets and a critical lack of standardized evaluation protocols in the field.

Motivated by the need to enhance efficiency, reduce costs, and establish a framework for rigorous, comparable research in geosciences, this work systematically analyzes current deep learning methods, architectures, and metrics. The proposed benchmark utilizes diverse public seismic datasets (F3 Netherlands, Penobscot, Parihaka) and standard performance metrics to provide a unified, objective, and reproducible evaluation framework for future seismic segmentation models.

References:

• Monteiro, B. A. A., Canguçu, G. L., Jorge, L. M. S., Vareto, R. H., Oliveira, B. S., Silva, T. H., Lima, L. A., Machado, A. M. C., Schwartz, W. R., & Vaz-de-Melo, P. O. S. (2024). Literature review on deep learning for the segmentation of seismic images. Earth-Science Reviews, 258, 104955. https://doi.org/10.1016/j.earscirev.2024.104955

• Canguçu, G. L., Jorge, L. M. S., Barreto, G. T., Silva, T. H., Lima, L. A., Caldas, W. S., Tavares, C. G. S., Schwartz, W. R., Vaz-de-Melo, P. O. S., & Machado, A. M. C. (n.d.). A Principled Benchmark for Seismic Data Segmentation.

2024

License Plate Recognition

This research addresses the critical challenge of License Plate Recognition (LPR) in real-world scenarios, where low-resolution (LR) and low-quality images from surveillance systems often lead to characters blending with the background or neighboring characters, severely hindering accurate recognition. Existing super-resolution methods often degrade network accuracy by failing to differentiate similar characters (e.g., 'Q' and 'O', 'B' and '8') or focus solely on objective metrics like PSNR and SSIM, which do not correlate well with actual LPR performance. Furthermore, a significant problem is the scarcity of publicly available datasets with paired real-world LR and high-resolution (HR) LP images and standardized evaluation protocols, leading to reliance on private or synthetically degraded data that fails to replicate real-world noise patterns.

Motivated by the urgent need for robust, efficient, and accurate LPR systems for applications like traffic monitoring and forensic investigations, this work proposes novel super-resolution approaches. These include new perceptual loss functions (e.g., Layout and Character Oriented Focal Loss - LCOFL) that integrate character recognition and LP layout awareness, enhanced architectures utilizing PixelShuffle layers, deformable convolutions, and attention modules for improved feature extraction and reorganization, and the introduction of UFPR-SR-Plates, a new public dataset and benchmark using real-world data and multi-image fusion strategies to provide a more realistic and rigorous evaluation framework.

References:

• Nascimento, V., Laroca, R., Lambert, J. d. A., Schwartz, W. R., & Menotti, D. (2022). Combining attention module and pixel shuffle for license plate super-resolution. Conference on Graphics, Patterns and Images (SIBGRAPI), 228–233. https://doi.org/10.1109/SIBGRAPI55357.2022.9991753

• Nascimento, V., Laroca, R., Ribeiro, R. O., Schwartz, W. R., & Menotti, D. (2024). Enhancing license plate super-resolution: A layout-aware and character-driven approach. Conference on Graphics, Patterns and Images (SIBGRAPI) 2024. https://doi.org/10.1109/SIBGRAPI62404.2024.10716303

• Nascimento, V., Lima, G. E., Ribeiro, R. O., Schwartz, W. R., Laroca, R., & Menotti, D. (2025). Toward Advancing License Plate Super-Resolution in Real-World Scenarios: A Dataset and Benchmark. Journal of the Brazilian Computer Society. https://valfride.github.io/nascimento2024toward/

• Nascimento, V., Laroca, R., Lambert, J. d. A., Schwartz, W. R., & Menotti, D. (2023). Super-Resolution of License Plate Images Using Attention Modules and Sub-Pixel Convolution Layers. Computers & Graphics, 113, 69–76. https://doi.org/10.1016/j.cag.2023.05.005

• Gonçalves, G. R., Diniz, M. A., Laroca, R., Menotti, D., & Schwartz, W. R. (2019). Multi-Task Learning for Low-Resolution License Plate Recognition. In Iberoamerican Congress on Pattern Recognition (CIARP), 251–261. https://doi.org/10.1007/978-3-030-33904-3_23

2021

Deep Learning Model Pruning

This research focuses on Deep Network Compression for Convolutional Neural Networks (CNNs), tackling the critical problem of their high computational cost, energy consumption, and memory requirements that hinder deployment on resource-constrained systems like mobile and IoT devices. A key motivation is to overcome limitations of existing pruning methods that often degrade network accuracy by indiscriminately removing filters or demand extensive effort.

The work proposes novel discriminative structured pruning strategies, including filter pruning and layer pruning, both leveraging Partial Least Squares (PLS) to identify and remove less important components while preserving or improving network accuracy. This aims to achieve superior resource-efficiency, including reduced FLOPs, memory usage, and prediction time, by intelligently reducing network width and depth.

References:
• Jordao, A., Yamada, F., & Schwartz, W. R. (2020). Deep Network Compression based on Partial Least Squares. Neurocomputing.
• Jordao, A., Lie, M., & Schwartz, W. R. Discriminative Layer Pruning for Convolutional Neural Networks