This study presents a novel dynamic depth estimation approach integrated with semantic fusion to improve visual perception accuracy in foggy conditions for autonomous vehicle systems in real-world driving environments.

Oscar Matthew Delgado, Tian Zhaowei, Meera Anil Menon, Philippe René Dubois, Alicia Karen Novak, Hiroki Satoshi Aida

This research enhances traffic scene analysis by combining multiscale attention networks with road contextual encoding, significantly boosting the accuracy of lane and sign detection for autonomous navigation systems.

Leonard James Whitmore, Gao Xiaojun, Nisha Kiran Devraj, Tobias Emanuel Kuhn, Alessia Maria Pellegrini, Wu Fanrong

This paper proposes a context-aware visual model for estimating crowd density in smart cities using multi-resolution attention maps, enabling adaptive monitoring during large-scale urban events and emergency responses.

Isabelle Nora Bennett, Chen Yuhao, Vikram Ganesh Shastry, Kai Tetsuo Nakamura, Claudia Elise Zimmermann, Omar Youssef El-Baroudi

This paper introduces an attention-augmented 3D convolutional neural network for recognizing human gestures in collaborative robot applications, ensuring efficient and safe task communication through intuitive body movement interpretation.

Carlos Eduardo Marin, Huang Lei, Tanvi Gaurav Nambiar, Michael Johann Bauer, Yu Yiqing, Petra Nicole Schneider

This research enhances the robustness of object detection systems for autonomous vehicles by applying adversarial training techniques under varied environmental perturbations, ensuring consistent performance across unpredictable driving conditions.

Benjamin Lucas Thornton, Li Chenghao, Ayesha Rani Kapadia, Marco Antonio Gutiérrez, Haruto Kenji Matsumoto, Elina Sophia McCarthy