This study presents a cross-view action recognition framework that leverages graph-based relational attention with spatial context encoding, improving accuracy in multi-camera surveillance scenarios with varied viewpoints and occlusions.

Frederick Simon Langley, Huang Yaqing, Richa Kumari Deshmukh, Mateo Ignacio Paredes, Charlotte Isabelle Leclerc, Zhou Xinming

This paper develops a robust aerial navigation framework for drones in GPS-denied zones, using visual feature matching and global pose estimation to ensure precise trajectory control in unknown terrains.

Theodore Miles Hamilton, Zhang Yunqi, Shalini Vijayrao Patil, Lucas Raphael Boudreaux, Hanae Rin Nakamura, Omar Khaled Bassem

This work introduces a fine-grained material classification model for robotic grasping tasks using texture-aware attention layers in vision transformers, enabling adaptive and accurate manipulation of diverse surface types.

Quentin Louis Bernard, Xu Wenhao, Meenal Sarita Krishnan, Daniel James Hollister, Yuki Noriko Takahashi, Sofia Isabella Müller

This study presents a real-time visual coordination framework for multi-agent systems enabling seamless object handover in human-robot collaborative warehouses through synchronized detection, gesture parsing, and movement prediction models.

Julian Patrick Whitaker, Liu Zihan, Kavya Ramesh Nair, Thomas Philippe Verdier, Gabriela Adriana Montoya, Zhang Chengyu

This paper presents a 3D scene reconstruction pipeline from egocentric video footage for augmented reality tools, aiding industrial maintenance tasks with spatial overlays and real-time component identification assistance.

Leonardo Emiliano Costa, Zhang Xinyuan, Poonam Sheetal Rathi, Benjamin Charles Müller, He Wenjing, Natalia Sofía Pereira

This study proposes a hierarchical contextual feature pyramid network to improve pedestrian detection in crowded urban environments, enhancing visibility under severe occlusions and varying human poses using refined attention layers.

Isaac Nathaniel Carmichael, Zhang Yuwei, Priyanka Deepak Nambiar, Louis Marcel Fontaine, Jin Minghao, Helena Sofia Duarte

This research presents an adaptive visual place recognition system using illumination-invariant embeddings, enabling robust localization for long-term autonomous navigation across drastically changing lighting conditions and seasons.

Theodore Julian Prescott, Li Xiaowen, Anjali Mohan Iyer, Fabrizio Marco Leone, Yukiko Haruna Watanabe, Clara Isabelle Schubert