[Paper Review] Non-rigid Registration Method between 3D CT Liver Data and 2D Ultrasonic Images based on Demons Model

Shuo Huang, Ke Wu+2 more|arXiv (Cornell University)|Dec 31, 2019

Medical Image Segmentation Techniques51 references4 citations

TL;DR

This paper proposes a non-rigid registration method between 3D CT liver data and 2D ultrasonic images using an enhanced Demons model. It introduces a boundary extraction and enhancement technique based on radial directional local intuitionistic fuzzy entropy in polar coordinates to improve registration accuracy, achieving sub-30-second processing time and superior performance over original Demons and Field II-simulated methods.

ABSTRACT

The non-rigid registration between CT data and ultrasonic images of liver can facilitate the diagnosis and treatment, which has been widely studied in recent years. To improve the registration accuracy of the Demons model on the non-rigid registration between 3D CT liver data and 2D ultrasonic images, a novel boundary extraction and enhancement method based on radial directional local intuitionistic fuzzy entropy in the polar coordinates has been put forward, and a new registration workflow has been provided. Experiments show that our method can acquire high-accuracy registration results. Experiments also show that the accuracy of the results of our method is higher than that of the original Demons method and the Demons method using simulated ultrasonic image by Field II. The operation time of our registration workflow is about 30 seconds, and it can be used in the surgery.

Motivation & Objective

• Improve the accuracy of non-rigid registration between 3D CT liver data and 2D ultrasonic images, which is challenging due to low contrast and noise in ultrasound.
• Address the limitations of the standard Demons model in handling low signal-to-noise ratio and poor boundary visibility in ultrasound images.
• Develop a novel preprocessing method to enhance image boundaries using radial directional local intuitionistic fuzzy entropy in polar coordinates.
• Create a complete registration workflow that integrates boundary enhancement with the Demons algorithm for improved convergence and robustness.
• Achieve clinically feasible processing time (under 30 seconds) while outperforming existing methods in registration accuracy.

Proposed method

• Propose a boundary enhancement method based on radial directional local intuitionistic fuzzy entropy in polar coordinates to amplify weak tissue boundaries in 2D ultrasound images.
• Apply a Gaussian filter to regularize the displacement field during iterative Demons optimization, ensuring smooth deformation fields.
• Use normalized mutual information (NMI) as the similarity measure to improve robustness in multi-modal registration between CT and ultrasound.
• Integrate the enhanced ultrasound images into the Demons algorithm, which uses a velocity field equation to estimate pixel-wise displacement based on intensity differences.
• Implement a relaxation factor and maximum iteration control to stabilize convergence and prevent overfitting.
• Combine BM4D-based denoising of 3D CT data before registration to improve input quality and reduce noise interference.

Experimental results

Research questions

• RQ1 Can radial directional local intuitionistic fuzzy entropy in polar coordinates effectively enhance weak boundaries in 2D ultrasound images for improved registration?
• RQ2 Does the proposed boundary enhancement method significantly improve the accuracy of non-rigid Demons-based registration between 3D CT and 2D ultrasound liver data?
• RQ3 How does the proposed method compare to the original Demons model and the Field II-simulated ultrasound method in terms of registration accuracy and computational time?
• RQ4 Can the proposed workflow achieve sub-30-second processing time suitable for real-time surgical applications?
• RQ5 To what extent does boundary enhancement reduce the risk of local minima in the optimization process during non-rigid registration?

Key findings

• The proposed method achieves higher registration accuracy than both the original Demons model and the Demons method using Field II-simulated ultrasound images.
• The average operation time of the proposed registration workflow is approximately 30 seconds, making it suitable for clinical use in surgery.
• The boundary enhancement technique based on radial directional local intuitionistic fuzzy entropy significantly improves the visibility of organ and tissue boundaries in low-contrast ultrasound images.
• The use of normalized mutual information (NMI) as the similarity measure enhances robustness in multi-modal registration, especially under intensity variations.
• The integration of BM4D denoising for 3D CT data improves the quality of input data and contributes to more accurate deformation fields.
• The method demonstrates strong anti-noise performance and effective handling of non-smooth deformations, confirming the suitability of the Demons model with enhancements for abdominal image registration.