A real-time, robust and versatile visual-SLAM framework based on deep learning networks

Zhang Xiao, Shuaixin Li

This paper explores how deep learning techniques can improve visual-based SLAM performance in challenging environments. By combining deep feature extraction and deep matching metho…