Alberto F. De Souza

Universidade Federal do Espírito Santo

Papers

19

Total Citations

1,373

H-Index

12

About

Alberto F. De Souza is a Brazilian computer scientist and engineer whose research spans autonomous vehicles, computer vision, deep learning, and brain-computer interfaces — a breadth that reflects his commitment to applying intelligent systems to real-world challenges. He is perhaps best known for his influential 2016 work on facial expression recognition using Convolutional Neural Networks, which has accumulated nearly 800 citations and addressed the critical practical challenge of training robust models with limited data. His contributions to autonomous driving have been particularly substantial: he has advanced large-scale environmental mapping, GNSS-denied localization, traffic light recognition using deep learning with prior maps, and obstacle avoidance systems through the IARA autonomous car project. His 2020 self-driving cars survey further cemented his role as a synthesizer of knowledge in this rapidly evolving field. De Souza has also made meaningful contributions to robotics logistics, developing RGBD-based datasets for warehouse pick-and-place operations. More recently, he has expanded into neurotechnology, publishing notable work on EEG-based motor imagery classification and lower-limb kinematic decoding during pedaling tasks. His diverse, application-driven portfolio — with over 1,200 cumulative citations — marks him as a versatile and impactful researcher bridging multiple engineering disciplines.

Research Focus

Key Achievements

12
H-Index
19
Papers
1,373
Total Citations
72
Avg Citations/Paper
🏆 Most Cited Paper
Facial expression recognition with Convolutional Neural Networks: Coping with few data and the training sample order
798 citations · 2016
📈 Most Prolific Year: 2016 (6 Papers)
🤝 Key Collaborators: 43
🏛 Institutions: Universidade Federal do Espírito Santo

Top Papers

  1. 1
  2. 2
  3. 3
  4. 4
  5. 5
    Self-driving cars: A survey
    53 citations · 2020
  6. 6
  7. 7
  8. 8
  9. 9
  10. 10

Key Collaborators

Contact & Links

Available for collaboration
Content generated · 14 days ago