Rigid registration algorithm based on the minimization of the total variation of the difference map

Area of Science:

• Medical Imaging
• Computer Vision
• Image Processing

Background:

• Image registration aligns similar scenes across different images.
• Challenges arise with significant variations in image contrast and background.
• Existing methods struggle with these dissimilarities.

Purpose of the Study:

• To propose a novel dissimilarity metric for rigid image registration.
• To develop and implement a method based on Total Variation of the Difference Map (TVDM) minimization.
• To enhance the accuracy and robustness of image registration for challenging datasets.

Main Methods:

• Utilized Total Variation of the Difference Map (TVDM) as a dissimilarity metric.
• Implemented a minimization-based algorithm for rigid image registration.
• Tested the method on synthesized and real experimental data with diverse noise and background conditions.

Main Results:

• The proposed TVDM-based method demonstrated high accuracy and robustness.
• Outperformed other rigid registration methods across all tested conditions.
• Successfully aligned images with vastly different contrasts and backgrounds.

Conclusions:

• The TVDM minimization method offers a robust solution for rigid image registration.
• This algorithm is particularly valuable for nano-scale X-ray imaging and microscopy.
• Provides a reliable option for critical image alignment applications.