About me

Hello, my name is André Luiz Buarque Vieira e Silva and I am a PhD Student in Computer Science at Centro de Informática (CIn) in Universidade Federal de Pernambuco (UFPE). I have experience in the topics of Computer Vision, Deep Learning, Particle-based Fluid Simulation and CUDA parallelization. My current research interest is in Image-based Industrial Inspection and Anomaly Detection in the wild. Facial biometrics is also a current interest.

My PhD research focus is to propose and investigate semi-supervised Deep Learning approaches to improve automatic defect and anomaly detection on data from “in-the-wild” scenarios, in which defect-related data is quite rare (for example, UAV/drone images of power line assets). Through new datasets (see InsPLAD) I am exploring the entire inspection process, from the data collection step until deploying the entire pipeline into production. Again, this is a challenging scenario since there is little or no prior information regarding the anomaly/defect itself.

During 2022 I was in Germany at TU Chemnitz as a DAAD scholarship holder for six months to deepen my PhD studies. I worked under the guidance of Jun.-Prof. Danny Kowerko in the Media Computing research group to advance the state of the art in Attention-based Image-level Anomaly Detection. From this experience, in 2023, we had an accepted poster in a CVPR workshop — 1st workshop on Vision-based InduStrial InspectiON (VISION) and a WACV 2024 main conference paper: AttentDifferNet.

Also in 2023, we had a dataset paper, InsPLAD, published at the International Journal of Remote Sensing (Impact Factor: 3.4). This paper was a result of my PhD research in collaboration with two brazilian companies: In Forma Software and Sistema de Transmissão Nordeste.

My curriculum vitae is available here.