Deep Learning-Powered Visual SLAM Aimed at Assisting Visually Impaired Navigation

About

Despite advancements in SLAM technologies, robust operation under challenging conditions such as low-texture, motion-blur, or challenging lighting remains an open challenge. Such conditions are common in applications such as assistive navigation for the visually impaired. These challenges undermine localization accuracy and tracking stability, reducing navigation reliability and safety. To overcome these limitations, we present SELM-SLAM3, a deep learning-enhanced visual SLAM framework that integrates SuperPoint and LightGlue for robust feature extraction and matching. We evaluated our framework using TUM RGB-D, ICL-NUIM, and TartanAir datasets, which feature diverse and challenging scenarios. SELM-SLAM3 outperforms conventional ORB-SLAM3 by an average of 87.84% and exceeds state-of-the-art RGB-D SLAM systems by 36.77%. Our framework demonstrates enhanced performance under challenging conditions, such as low-texture scenes and fast motion, providing a reliable platform for developing navigation aids for the visually impaired.

Marziyeh Bamdad, Hans-Peter Hutter, Alireza Darvishy · 2025

Related benchmarks

Task	Dataset	Result
Visual SLAM	TUM RGB-D fr1 desk	ATE RMSE (cm)1.9	24
Trajectory Estimation	TUM RGB-D Freiburg1	RMSE0.019	17
Visual SLAM	TUM RGB-D fr1/room	Translation RMSE (m)0.138	8
Visual SLAM	TUM RGB-D fr1/desk2	ATE RMSE0.04	7
SLAM	TartanAir Hospital-Hard sequences	ATE (m) P0370.049	4
SLAM	TUM RGB-D Freiburg1 plant	ATE (m)0.034	3
Visual SLAM	ICL-NUIM (various sequences)	ATE (m) - lr-kt00.006	3
SLAM	TUM RGB-D Freiburg1 rpy	ATE (m)0.021	2
SLAM	TUM RGB-D Freiburg1 teddy	ATE (m)0.131	2
SLAM	TUM RGB-D Freiburg1 floor	ATE (m)0.04	2