Byron Leite Dantas Bezerra
Papers
7
Total Citations
108
H-Index
5
About
Byron Leite Dantas Bezerra is a leading researcher at the intersection of computer vision, deep learning, and robotics, with a primary focus on handwritten and scene text recognition. His most influential work, "HTR-Flor: A Deep Learning System for Offline Handwritten Text Recognition" (2020, 57 citations), advances state-of-the-art Convolutional Recurrent Neural Networks (CRNNs) for this challenging domain. Bezerra has also made significant contributions to developing compact, efficient neural architectures for real-world applications, including octave convolutional networks for fire recognition (2019, 13 citations) and multilingual text detection (2019, 11 citations). His research extends into socially assistive robotics, where he systematically analyzed computer vision's role (2022, 12 citations), and into medical training, optimizing CNNs for robotic surgery skill evaluation (2019, 11 citations). Notably, his work on the NAO-Read system empowers humanoid robots to recognize text in natural scenes, bridging robotics and document analysis. With a portfolio addressing occlusion detection in document scanning and lightweight models for hardware-constrained devices, Bezerra's research demonstrates a consistent commitment to deploying robust, efficient AI solutions across diverse, high-impact applications.
Research Focus
Key Achievements
Top Papers
- 1HTR-Flor: A Deep Learning System for Offline Handwritten Text Recognition57 citations · 2020
- 2
- 3
- 4
- 5
- 6
- 7