Mateusz Dziwulski

Papers

Total Citations

H-Index

About

Mateusz Dziwulski is a researcher advancing the frontiers of autonomous navigation through deep learning, with a primary focus on visual odometry and its applications in robotics and automated driving. His most-cited work, "Weakly Supervised End2End Deep Visual Odometry" (2024), tackles the ill-posed problem of estimating camera motion from visual data, demonstrating that deep models can surpass traditional approaches in localization accuracy while dramatically reducing catastrophic failures. This contribution is particularly impactful for mapless navigation systems, where robust, end-to-end learning methods are essential. Dziwulski’s research bridges the gap between theoretical computer vision and practical robotics, offering scalable solutions for real-world deployment. With 2 citations already for this recent paper, his work is gaining traction among peers seeking more reliable and efficient odometry pipelines. By leveraging weakly supervised techniques, he reduces reliance on costly labeled data, making his methods more accessible for diverse autonomous systems. Dziwulski’s innovative approach positions him as a promising voice in the field, with potential to shape future developments in self-driving technology and mobile robotics.