A novel enhanced intensity-based automatic registration: Augmented reality for visualization and localization cancer...

Area of Science:

• Medical Imaging
• Computer-Aided Surgery
• Image Registration

Background:

• Manual landmark selection for mesh and video registration is time-consuming and prone to inaccuracies.
• Accurate real-time visualization of tumors is crucial for effective surgical guidance.
• Existing methods lack the precision and speed required for seamless augmented reality integration.

Purpose of the Study:

• To develop and evaluate an Intensity-based Automatic Registration (IbAR) method to replace manual processes.
• To enhance registration accuracy and reduce processing time for augmented reality in surgical tumor visualization.
• To introduce the Enhanced Intensity-based Automatic Registration (EIbAR) system utilizing a Modified Zero Normalized Cross Correlation (MZNCC) algorithm.

Main Methods:

• Implementation of the Enhanced Intensity-based Automatic Registration (EIbAR) system.
• Utilized the Modified Zero Normalized Cross Correlation (MZNCC) algorithm for image registration.
• Tested the system on videos of breast cancer tumors for scene augmentation.

Main Results:

• The EIbAR system achieved an average improvement in registration accuracy of 2 mm compared to a reference method.
• Reduced pixel matching, leading to an average decrease in processing time of 22 ms/frame.
• Demonstrated acceptable accuracy and processing time for real-time surgical applications.

Conclusions:

• The proposed EIbAR system offers a significant advancement over manual registration techniques.
• The method provides a viable solution for accurate and efficient tumor visualization in augmented reality surgery.
• Enables seamless integration of augmented reality for surgeons, enhancing cancer tumor visualization during procedures.