4 papers
MuViSeg: Multi-View Segment Correspondences from Dense Geometry Priors
Denis Fatykhoph, Timur Akhtyamov, Konstantin Pakulev +2
Classical image correspondence is solved at the level of sparse keypoints or dense pixels, but the systems that consume these matches - object-level mapping, topological navigation…
EgoWalk: A Multimodal Dataset for Robot Navigation in the Wild
Timur Akhtyamov, Mohamad Al Mdfaa, Javier Antonio Ramirez Benavides +11
Data-driven navigation algorithms are critically dependent on large-scale, high-quality real-world data collection for successful training and robust performance in realistic and u…
Spatiotemporal Knowledge Graphs as Persistent Scene Memory for Embodied Question Answering
Mohamad Al Mdfaa, Svetlana Lukina, Timur Akhtyamov +4
Vision-language models (VLMs) demonstrate strong image-level scene understanding, but reasoning over long egocentric video remains costly: because VLMs maintain no persistent memor…
PixelNav: Towards Model-based Vision-Only Navigation with Topological Graphs
Sergey Bakulin, Timur Akhtyamov, Denis Fatykhov +2
This work proposes a novel hybrid approach for vision-only navigation of mobile robots, which combines advances of both deep learning approaches and classical model-based planning…