learning and feeling SLAM together with hands-on-experiments
-
Updated
Sep 5, 2021 - Jupyter Notebook
learning and feeling SLAM together with hands-on-experiments
BodySLAM is a robust deep learning-based Simultaneous Localization and Mapping (SLAM) approach designed for endoscopic surgical applications. This framework effectively operates across various surgical settings, including laparoscopy, gastroscopy, and colonoscopy.
To associate your repository with the python-slam topic, visit your repo's landing page and select "manage topics."