Zakaria Laskar

Papers

Total Citations

H-Index

About

Zakaria Laskar is a computer vision researcher whose work centers on visual localization, scene understanding, and deep learning for robotics. His most influential contribution is HSCNet++, a hierarchical framework that combines scene coordinate classification and regression using Transformer architectures, achieving state-of-the-art performance in single-image RGB localization. This work, published in 2024 with 23 citations, addresses the critical challenge of enabling robots and autonomous systems to determine their position from a single photograph without relying on expensive 3D models or depth sensors. Laskar's research bridges the gap between traditional feature-based methods and modern neural approaches, making visual localization more robust and efficient for real-world deployment. His innovations in hierarchical scene coordinate prediction have significant implications for augmented reality, autonomous navigation, and mobile robotics. By integrating Transformer networks into the localization pipeline, Laskar has demonstrated how attention mechanisms can improve accuracy in challenging environments with varying lighting, viewpoints, and occlusions. His ongoing work continues to push the boundaries of what is possible in visual localization, establishing him as an emerging leader in this critical area of computer vision.