Aleksander Byrski

Papers

Total Citations

H-Index

About

Aleksander Byrski is a researcher advancing the frontiers of computer vision and multi-agent robotics, with a focus on human pose estimation and decentralized coordination. His most cited work, "Combined YOLOv5 and HRNet for High Accuracy 2D Keypoint and Human Pose Estimation" (2022, 20 citations), introduces a novel hybrid architecture that integrates YOLOv5’s object detection speed with HRNet’s high-resolution feature extraction. This approach significantly improves the precision of 2D keypoint localization, enabling robust human pose tracking in challenging real-world scenarios such as sports analysis, medical fall detection, and human-robot interaction. Byrski’s contributions address critical limitations in CNN-based pose estimation, offering a scalable solution for dynamic environments. Additionally, his work on "Socially-inspired fully decentralized robot coordination" (2022) explores biologically and socially inspired algorithms for multi-robot systems, emphasizing autonomy and scalability without centralized control. Byrski’s research bridges computer vision and swarm robotics, demonstrating practical impact in both fields. His innovative methods continue to influence applications requiring real-time, accurate human-robot interaction and collaborative autonomous systems.