Bug: Inaccurate Registration Due to Lack of Moving Mask During Image Alignment

[LalithShiyam]

Summary

The current CT image alignment tool is experiencing inaccurate registration when aligning CT images of the same patient acquired at different time points. This issue seems to stem from the tool only considering the fixed mask, leading to the moving image being improperly aligned.

Problem Description

The algorithm is designed to align a pair of CT images (fixed and moving). However, without incorporating the moving mask into the registration process, the moving image is inaccurately dragged toward the fixed mask. This results in a misalignment that could significantly impact clinical decision-making.

Expected Behavior

The tool should consider both the fixed mask and the moving mask during the registration process to achieve an accurate overlay of the CT images.

Steps to Reproduce

1. Load the fixed and moving CT images into the tool.
2. Apply the fixed mask to the moving image using the current alignment algorithm.
3. Observe the misalignment in the registered images.

Proposed Solution

Modify the image registration class to accept and utilize both the fixed mask and the moving mask. This should provide a more accurate reference for alignment and prevent the dragging issue.

Acceptance Criteria

• The registration algorithm successfully aligns the moving image to the fixed image using both masks.
• The accuracy of the overlay is verified by comparing pre- and post-registration overlays.
• Clinical efficacy is confirmed by domain expert review.

[LalithShiyam]

This is not relevant anymore as the code logic is very different.