Luiz Gustavo Hafemann

Luiz Gustavo Hafemann

Computer Vision Researcher

About Me

Computer Vision Researcher at SPORTLOGiQ, applying machine learning and computer vision to sports analytics. I have been modelling and training deep learning systems for computer vision for the past 6 years, with 9 published articles, including top tier venues such as CVPR and IEEE Transactions on Information Forensic and Security..

I obtained my Ph.D. from ÉTS Montreal in 2019. My research was focused on the biometric task of Offline Handwritten Signature Verification. I also have a master's in Computer Science, in which I explored Deep Learning models for Texture Classification. I participated in the 2017 and 2018 NIPS competition on adversarial attacks/defenses, and my team got 1st place on the "Untargeted attacks" track in the 2018 edition. I also participated in two editions of the Deep Learning Summer School organized by the MILA lab (in 2015 and 2016).

Download Resume

Publications

Journal Papers

Hafemann, L., Sabourin, R., Oliveira, L. "Characterizing and evaluating adversarial examples for Offline Handwritten Signature Verification", IEEE Transactions on Information Forensics and Security , 2019. paper, preprint code

Hafemann, L., Sabourin, R., Oliveira, L. "Fixed-sized representation learning from Offline Handwritten Signatures of different sizes", International Journal on Document Analysis and Recognition (IJDAR), 2018. paper, preprint code

Hafemann, L., Sabourin, R., Oliveira, L. "Learning Features for Offline Handwritten Signature Verification using Deep Convolutional Neural Networks", Pattern Recognition, 2017. paper, preprint code

Conference Papers

Rony, J.*, Hafemann, L.*, Oliveira, L., Ben Ayed, I., Sabourin, R., Granger, E. "Decoupling Direction and Norm for Efficient Gradient-Based L2 Adversarial Attacks and Defenses", Conference on Computer Vision and Pattern Recognition (CVPR oral presentation), 2019. link code

Hafemann, L., Sabourin, R., Oliveira, L., "Offline Handwritten Signature Verification - Literature Review", International Conference on Image Processing Theory, Tools and Applications (IPTA), 2017. link

Hafemann, L., Oliveira, L., Sabourin, R. "Analyzing features learned for Offline Signature Verification using Deep CNNs", International Conference on Pattern Recognition (ICPR), 2016. link

Hafemann, L., Oliveira, L., Sabourin, R. "Writer-independent Feature Learning for Offline Signature Verification using Deep Convolutional Neural Networks", International Joint Conference on Neural Networks (IJCNN), 2016. link

Hafemann, L., Cavalin, P., Oliveira, L., Sabourin, R. "Transferring Learning between Texture Classification Tasks using Convolutional Neural Networks", International Joint Conference on Neural Networks (IJCNN), 2015. link

Hafemann, L., Cavalin, P., Oliveira, L. "Forest Species Recognition using Deep Convolutional Neural Networks". 22nd International Conference on Pattern Recognition (ICPR), 2014. link

* Equal contribution

Thesis

Hafemann, L. "Learning features for offline handwritten signature verification", Ph.D. thesis, Systems Engineering Department, École de technologie supérieure, 2019. link, pdf

Hafemann, L. "An Analysis of Deep Neural Networks for Texture Classification", Master's thesis, Department of Informatics, Federal University of Paraná, 2014. link, pdf

Preprints

Abdoli, S., Hafemann, L., Rony, J., Ben Ayed, I., Cardinal, P., Koerich, A. "Universal Adversarial Audio Perturbations" preprint

Cruz, R., Hafemann, L., Sabourin, R., Cavalcanti, G. "DESlib: A Dynamic ensemble selection library in Python" preprint code