Mihai Lorin Dimoiu
Universitatea Națională de Știință și Tehnologie Politehnica București
Papers
1
Total Citations
2
H-Index
1
About
Mihai Lorin Dimoiu is a researcher specializing in computer vision and deep learning, with a particular focus on generative models and their application to remote sensing. His work centers on advancing image segmentation techniques for aerial imagery, a critical area for autonomous systems, urban planning, and environmental monitoring. Dimoiu’s most notable contribution is his 2021 paper, "Improved Conditional GAN for Aerial Image Segmentation," which explores how Generative Adversarial Networks (GANs)—architectures where two neural networks compete to generate realistic synthetic data—can be optimized for pixel-level classification of aerial scenes. By systematically comparing different GAN implementations, his research provides practical insights into enhancing segmentation accuracy, addressing challenges like boundary delineation and class imbalance in complex overhead views. Though early in his career, with his key work accumulating 2 citations, Dimoiu’s investigation into adversarial training for geospatial data lays a foundation for more robust and automated interpretation of satellite and drone imagery. His contributions are particularly relevant for researchers seeking to bridge the gap between generative modeling and high-stakes visual understanding tasks.
Research Focus
Key Achievements
Top Papers
- 1Improved Conditional GAN for Aerial Image Segmentation2 citations · 2021