| MS Seminar


Name of the Speaker: Mr. Abdul Wahid (EE21S053)
Guide: Dr. Mohanasankar S
Online meeting link: http://meet.google.com/zdj-wpnm-abf
Date/Time: 12th May 2025 (Monday), 3:00 PM
Title: Stereo Vision-Based Markerless Registration Method for Image-Guided Cranial Surgery.

Abstract :

Image-guided cranial surgery depends on accurate registration of preoperative medical images with the patient’s anatomy to enable precise navigation during complex procedures. Traditional registration methods, which rely on manual landmark selection or fiducial markers, are time-consuming, prone to human error, and may increase infection risks. This seminar presents a stereo vision-based markerless registration method that integrates machine vision and deep learning to enhance the precision and efficiency of image-guided cranial surgery. Using a stereo camera setup, we evaluate both classical stereo reconstruction techniques and a deep learning-based (LEAStereo) based approach for 3D surface reconstruction of a head phantom. Our proposed method automatically detects anatomical landmarks and aligns them with preoperative CT scans, achieving a mean target registration error of 0.973 mm—outperforming manual methods. By eliminating physical markers, this method streamlines surgical workflows and enhances patient safety. The seminar will explore the methodology, experimental results, and potential for clinical integration, advancing the future of precision surgical navigation.